Computers in Human Behavior 101 (2019) 352–370

Contents lists available at ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier.com/locate/comphumbeh

Mobile crowd sensing – Taxonomy, applications, challenges, and solutions T

a b,∗ c d
Djallel Eddine Boubiche , Muhammad Imran , Aneela Maqsood , Muhammad Shoaib
a
LaSTIC Laboratory, Department of Computer Science, University of Batna 2, Algeria
b
College of Applied Computer Science, King Saud University, Saudi Arabia
c
Department of Behavioral Sciences, Fatima Jinnah Women University, Rawalpindi, Pakistan
d
College of Computer and Information Sciences, King Saud University, Saudi Arabia

ARTICLE INFO ABSTRACT

Keywords: Recently, mobile crowd sensing (MCS) is captivating growing attention because of their suitability for enormous
Mobile crowd sensing range of new types of context-aware applications and services. This is attributed to the fact that modern
Multifacted infrastructural and human- smartphones are equipped with unprecedented sensing, computing, and communication capabilities that allow
powered applications them to perform more complex tasks besides their inherent calling features. Despite a number of merits, MCS
Social and behavioral applications
confronts new challenges due to network dynamics, the huge volume of data, sensing task coordination, and the
Large-scale sensing
Communication
user privacy problems. In this paper, a comprehensive review of MCS is presented. First, we highlight the dis-
Computing tinguishing features and potential advantages of MCS compared to conventional sensor networks. Then, a tax-
onomy of MCS is devised based on sensing scale, level of user involvement and responsiveness, sampling rate,
and underlying network infrastructure. Afterward, we categorize and classify prominent applications of MCS in
environmental, infrastructure, social, and behavioral domains. The core architecture of MCS is also described.
Finally, we describe the potential advantages, determine and reiterate the open research challenges of MCS and
illustrate possible solutions.

1. Introduction user-companioned devices (including mobile phones, smart vehicles,

In recent years, with the advancements in mobile and wireless
communication technologies (e.g., 4G/5G), efficient and high cap-
ability networks have emerged allowing better connectivity to billions
of ubiquitous smartphones and vehicular systems. Alongside the growth
of mobile networks, plethora of sensing devices have been integrated
into smartphones and vehicular systems enhancing thus their sensing
capability. Moreover, they are fitted out with more processing power
and storage capacities. This joint evolution led to an emergent para-
digm named Mobile Crowd Sensing (MCS).
MCS is an emerging sensing and geo-crowd sourcing paradigm,
based on the mobile device's sensor, that enables acquiring local
geospatial information and knowledge and giving the possibility to
share this information/knowledge with other users and wider com-
munity (Boulos et al., 2011; Ma, Zhao, & Yuan, 2014). MCS has been
also defined by Raghu Ganti as the data sharing and information ex-
traction process, based on the collaboration of individuals with sensing
and computing devices, to measure and enumerate mutual interest
phenomena (Ganti, Ye, & Lei, 2011). Other researchers also formally
define MCS as a large-scale sensing paradigm based on the power of
wearable devices, and so on) to contribute sensed or generated data
from their mobile devices and aggregate and fuse the data in the cloud
for crowd intelligence extraction and human centric service delivery
(Guo et al., 2014a).
Indeed, the urbanization of the world population and the fast
growth of smart cities (Mehmood, Ahmad, Yaqoob, Adnane, Imran,
Guizani) make MCS a promising research area that enables innovative
techniques and systems which exploit data and communications tech-
nologies in order to enhance their functions, improve efficiency, and
the economy and offer the best environment for their citizens. MCS
enables a plentiful number of individual mobile phones and vehicular
devices to share local knowledge (e.g., local information, ambient
context, noise level, and traffic conditions) collected by a myriad of
embedded sensors (e.g., GPS, digital compass, microphone, light sensor,
and accelerometer). The application of analysis, reasoning and data
mining techniques on the shared sensed data provides useful insights
for smart urban space monitoring and might have an enormous societal
impact.
Despite the recent emergence of the MCS concept, there are some
fruitful research works in the literature that addressed the crowd

Corresponding author.
∗

E-mail addresses: de.boubiche@univ-batna2.dz (D.E. Boubiche), dr.m.imran@ieee.org (M. Imran), dr_aneelamaqsood@fjwu.edu.pk (A. Maqsood),
muhshoaib@ksu.edu.sa (M. Shoaib).

https://doi.org/10.1016/j.chb.2018.10.028
Received 2 April 2018; Received in revised form 15 October 2018; Accepted 21 October 2018
Available online 22 October 2018
0747-5632/ © 2018 Elsevier Ltd. All rights reserved.
D.E. Boubiche et al.
sensing before. In (Mehmood et al., 2017), the authors introduced the
crowd wisdom concept which demonstrates that decisions made by a
group of people have better results than those made by a single person.
Four quality metrics have been proposed to define the efficiency of the
crowd: diversity in opinion, independence of thinking, decentralization,
and opinion aggregation (Surowiecki, 2004). Another notion close to
the crowd sensing concept, named participatory sensing, was proposed
in (Guo et al., 2015a). The main idea is to charge individual mobile
devices to build participatory sensor networks that enable collecting,
evaluating, and sharing local knowledge. Compared to the participatory
sensing which is only devoted to public sensing, MCS involves both
personal and public sensing. Moreover, explicit and implicit user par-
ticipation is considered in MCS while participatory sensing involves
only explicit user participation.
Mobile crowd sensing have been the subject of several review pa-
pers in recent years. Nicholas D. Lane et al. presented one of the first
survey papers on MCS, describing the existing mobile phone sensing
algorithms, applications, and systems (Lane et al., 2010). Furthermore,
the authors discussed the emerging sensing paradigms, and formulated
an architectural framework. Similarly, Khan, Wazir Zada et al. pre-
sented a comprehensive study on all the MCS systems which use smart
phones and mobile phone sensors for humans good and for better
human-phone interaction (Khan et al., 2013). Bin Guo et al. presented a
review study which characterizes the unique features and novel appli-
cation areas of MCS, and proposed a reference framework for building
human-in-the-loop MCS systems. Also, further clarification about the
complementary nature of human and machine intelligence have been
presented in (Guo et al., 2015a). Another paper that serves the MCS was
presented by Huadong Ma et al., where opportunistic characteristics of
human mobility were investigated from the perspectives of both sensing
and transmission (Ma et al., 2014). Also, the authors have discussed
how to exploit these opportunities to collect data efficiently and ef-
fectively.
The mobile crowd sensing remains an attractive research area and
numerous research works have been proposed recently. Indeed, the
application field continues to grow and several challenges and solutions
have emerged. To the best of our knowledge, none of the existing stu-
dies investigated various aspects of MCS which are explored and pre-
sented plored in this article. Therefore, it is very interesting to in-
vestigate the last emerging challenges and solutions for MCS and
highlight the introduced advantages and opportunities. The contribu-
tions of this study are summarized as follows:
• We highlight the distinguishing features of MCS compared to WSN
and describe its potential advantages.
• We present a taxonomy of MCS based on sensing scale, level of user
involvement and responsiveness, sampling rate, and network infra-
structure.
• We also categorize, classify, and briefly describe MCS emerging
applications into environmental, infrastructure, social, and beha-
vior.
• We describe the architecture of MCS based on the provided services
and the interacting entities.
• We present the main features and advantages of MCS and expose the
key architectures and the application domains.
• We identify, and reiterate the main challenges of MCS and outline
the current MCS strategies and solutions.
The paper is organized as follows: Section 2 highlights the distin-
guishing features and advantages of MCS compared to it's predecessor
i.e., wireless sensor network (WSN). Section 3 introduces a taxonomy of
the MCS. Section 4 presents an overview of the main MCS applications.
In section 5, we present a typical MCS architecture based on the pro-
vided services and the interacting entities. Section 6 points out the
current MCS strategies and solutions. Finally, we conclude the paper in
section 7.
353
Computers in Human Behavior 101 (2019) 352–370
2. MCS vs. WSN: unique features and benefits
Although MCS can be considered as an evolution of sensor network,
however, it has several unique characteristics and advantages com-
pared to traditional WSN.
2.1. Resource capacity
Unlike resource-constrained WSN, MCS is based mainly on mobile
phones, which are equipped with much more resources in terms of
computing power, memory storage, communication range, and energy
supply. Moreover, a rich set of sensors is integrated in mobile phones,
such as camera, GPS, accelerometer, microphone, gyroscope, while
traditional sensor devices are dedicated to sense limited sensing values.
Following the fast advances in smartphone technologies, innovative
smart devices, such as smart watch, smart glass, and smart clothes are
expected to be a part of MCS network enabling thus more insightful
applications and services.
2.2. Network deployment and coverage
The deployment of traditional WSN is often difficult and might be
expensive. On the other hand, MCS platforms are quite low cost and
there is no need to network deployment as billions of mobile phones are
already deployed in the field. Therefore, large-scale sensing networks
can be built with low cost in less time compared to WSN. The network
coverage represents another important MCS characteristic over WSN.
Indeed, the smart device users are often mobile which provides an
outstanding spatiotemporal coverage compared to traditional static
WSN deployments.
2.3. Hybrid sensing approach
MCS is mainly based on a hybrid sensing system where humans and
machines are both involved in the sensing task. Compared to WSN, MCS
offers a smart sensing process as it takes benefit of the human in-
telligence. Indeed, contributors can control the sensing process by de-
ciding what to be captured. Also, smart phones can capture con-
tributors' feelings or remarks regarding the readings, which optimizes
the sensing efficiency. Thus, without requiring sophisticated hardware
and software, higher quality or semantically complex data can be col-
lected by exploiting the intelligence and mobility of humans.
2.4. Heterogeneous network connectivity
Based on the MCS requirements and available network infra-
structures, mobile phones and devices are interconnected through dif-
ferent networking technologies. For example, GSM can be used for large
scale connectivity while Wi-Fi and Bluetooth can used for short-range
communications. Indeed, MCS enables an interconnection medium
which supports transient networking services, such as connectivity,
collaborative sensing, and data routing/transmission, for the partici-
pants crossing these multiple wireless networks (Surowiecki, 2004).
2.5. Network disruption tolerance
Due to the irregular network connectivity in some cases (e.g., low
signal strength due to interference or no signal in a rural area), efficient
links between mobile phones could not be guaranteed. This disrupted
network nature can be tolerated by most of MCS applications. Indeed,
real time transmission of individual sensed data is not always imposed
and MCS applications can tolerate also, in some cases, a low degree of
data accuracy. Mobile crowd applications enable disruption tolerance
services based on data storage strategies where captured data are stored
locally before their deliverance when the network connection is avail-
able.
D.E. Boubiche et al.
2.6. Wireless networking technologies
Several wireless networking technologies can be involved in the
mobile crowd sensing process. Indeed, the mobile crowd sensing par-
ticipants can communicate in short distances using short rang wireless
network technologies (e.g., Bluetooth, and Wifi). On the other hand, 5G
networks can be employed for long range communication of sensed
data from all kinds of mobile devices with various wireless interfaces.
2.7. Sensing platforms
MCS is made around several hybrid sensing platforms from classical
WSNs to newly widely held mobile devices such as: smartphones, ta-
blets, smart watches, smart glasses and smart bracelets. The fast growth
in internet of things and sensors technologies will lead probably to the
emergence of more mobile sensing platforms.
Indeed, the unique features of MCS enables plethora of novel ap-
plications compared to the classical WSN. The resource capacity of the
crowd sensing devices (GPS, accelerometer, microphone, gyroscope
etc.) can be exploited in variety of sensing applications. Moreover, the
networking connectivity, allows large scale deployment of MCS appli-
cations.
3. MCS taxonomy
Before surveying the main applications, challenges and solutions,
we devise a taxonomy of mobile crowd sensing based on the existing
literature. From scalability point of view, mobile sensing can be cate-
gorized as follows.
• Separate sensing: there is no collaboration in the sensing process,
and the data are gathered individually for personal use only.
• Cluster sensing: a group of users with common interests collaborate
and share the collected data.
• Community sensing: large call sensing based on peoples' commu-
nities which collaborate to predict the global trends.
Another taxonomy approach is based on the involvement level of
individuals that participate in the sensing process. In this taxonomy, we
can categorize the sensing approach as opportunistic, participatory, and
hybrid sensing (a mixture of user control and device processing). In the
opportunistic sensing, the data is collected in an automatic manner with
low user involvement. The participatory sensing operates in a different
way and requires the users participation in the data collection activity.
Also, the user responsiveness to the sensing process can be considered
as implicit or explicit.
From the viewpoint of the sampling rate, the data sensing can be
performed continuously or executed depending on a specific context
(certain time periods or places). However, the continuous sensing may
exhaust the sensing resources while the context aware sensing will be
inefficient to monitor dynamic events. Regarding the network infra-
structure, the MCS can exploit the existing infrastructures (e.g., access
points and GSM). Otherwise, the adhoc infrastructure can be adopted
forming thus opportunistic networks. Also, a hybrid infrastructure can
be used based on adhoc and existing network infrastructures (see
Fig. 1).
4. MCS application domains overview
MCS has opened several application opportunities in various do-
mains. These domains can be classified, as presented in Fig. 2, in four
main categories based on the type of sensed phenomenon: Environ-
mental, Infrastructure, social, and behavior (Ganti et al., 2011). In this
section, we review state-of-the-art MCS applications and projects.
354
Computers in Human Behavior 101 (2019) 352–370
Fig. 1. MCS taxonomy.
4.1. Environmental applications
The applications of MCS in the environmental field aims to basically
preserve the nature and monitor the space weather and air and noise
pollution levels. In the last decade, several scientific studies have been
conducted based on mobile devices of volunteer contributors.
• Space weather: a recent report by the National Research Council of
the National Academies (National Research Council of the National
Academies, 2012) has exposed the benefits of applying crowd sen-
sing for weather monitoring. To optimize the efficiency of air tem-
perature monitoring systems, user's smartphones are exploited to
establish a worldwide temperature sensor network at low cost and in
an automated fashion. Thus, there is no need for installing the ex-
pensive and limited area weather stations which moreover require
maintenance. One of the relevant examples of MCS application for
temperature measurement is the employing the Android application
Open Signal (http://opensignal.com/.) for smartphones. From 25
April 2012 to 22 April 2013 over 530 thousand devices have con-
tributed a total of over 220 million readings of weather tempera-
tures across more than 200 countries/territories (http://open-
signal.com/.). The Table 1 gives an overall picture of the most
known proposed research works and projects that exploit the notion
of MCS in weather monitoring field (see Tables 2–5).
• Air pollution: MCS can be applied to establish a mobile measure-
ment system which enables the generation of accurate ultrafine
particle pollution maps with high spatio-temporal resolution.
Volunteers carrying smartphones and sensors contribute the sensed
data while all citizens using the application can receive the mea-
surements from the data cloud in real-time and only for the areas of
their particular interest (Antonić et al., 2014). We further sum-
marize, in Table 1, the most representative pollution applications
based on MCS.
• Noise pollution: Noise pollution is as harmful as air and water
pollution, impacting the health and well being of citizens. MCS of-
fers new opportunities in the noise monitoring field, where users'
feelings or remarks regarding the readings can be collected by smart
phones. Citizens' participation represents an efficient approach of
sensing noise levels, as they can instantly sense it using their smart
phones and send it to the city officers if the noise is intolerable.
Several research works have addressed the MCS applications to
monitor noise pollution levels. Table 1 summarizes some of the
prominent applications of MCS for environmental monitoring.
D.E. Boubiche et al.
• Space
Environmental App
Infrastructure App
weather
• availability
• Air
•
• Noise
4.2. Infrastructure applications
Infrastructure monitoring represents a proliferate application field
for MCS, which consists of large scale measurement of phenomena re-
lated to public infrastructure involving traffic congestion, parking
availability, road conditions, real-time transit tracking (Liu, Liu,
Muhammad, Sangaiah, & Doctor, 2018), power line condition (Wang
et al., 2018) and outages of public works (e.g. Broken traffic lights or
malfunctioning fire hydrants).
Traffic analysis and transit tracking represent the most addressed
application fields in the infrastructure monitoring. One of the first re-
searches that introduce the use of the crowd sourcing concept in traffic
management was proposed by G.L. Graunke. A large amount of traffic
information is available from a variety of sources. This information may
be used by drivers to attempt to select from among a variety of potential
routes to travel to a given destination (Graunke, 2001). Nericell re-
presents another traffic congestion control system based on the use of
the accelerometer, microphone, GSM radio, and/or GPS sensors em-
bedded in user's phones to detect potholes, bumps, braking, and
honking (Mohan, Padmanabhan, & Ramjee, 2008). The traffic conges-
tion problem was also addressed in (Wen et al., 2008) where a MCS
system was proposed based on GPS-equipped taxis to analyze traffic
changes around the Olympic games in Beijing. The same idea was also
adopted in (Liu, Biderman, & Ratti, 2009), where floating GPS-
equipped car and smart card records from both bus and metro are used
for real time data evaluation of urban traffic. The application of MCS in
transit tracking was also introduced in (Thiagarajan et al., 2010) where
participating users install an application on their smartphones to ex-
ploit GPS, WiFi, and accelerometer for transit tracking. CY Lin et al.
proposed a novel Comfort Measuring System (CMS) for public trans-
portation systems where GPS and 3-axis accelerometer functions of
modern smart phones are used to measure the comfort level of the
vehicle rides (Lin et al., 2010).
To predict a bus arrival time, P. Zhou et al. proposed an MCS system
based on bus passengers' participatory sensing (Zhou, Zheng, & Li,
2012). SmartRoad is another traffic monitoring application which in-
troduces a road sensing system that generates and collects mobile
sensory data from vehicle-resident mobile phones (Hu et al., 2013).
TrafficInfo represents an MCS application which implements a
participatory sensing based live public transport information service
(Farkas et al., 2014). In (Bröring et al., 2015), the authors proposed
enviroCar which represents a new platform for collecting traffic and
related environmental data from automobile equipped with On-Board
Diagnostics (OBDII) and Android smartphone. J. Wan et al. proposed a
MCS technology to support the creation of dynamic route choices for
drivers wishing to avoid congestion (Wan, Liu, Shao, & et al, 2016). In
(Yan et al., 2017), the authors proposed a cloud-assisted MCS archi-
tecture for urban transportation system where traffic congestion is
Computers in Human Behavior 101 (2019) 352–370
• Traffic • Community
analysis behavior
• Travel assistance
• • Emergency
• Parking behavior
Social App
mobile phones
• • Parking discovery
• Road • Urban lifestyle
• Personal/Public and mobility
• health • Healthcare
• Transit • Urban social
tracking events/structure
Fig. 2. MCS applications' categories.
controlled based on sensed data obtained continuously from a large set
of smartphones carried by drivers. A. KAUR et al. proposed an efficient
and cost-effective method using smartphones to determine the traffic
state of the road (Kaur et al., 2017). RoadCrowd represents a novel
traffic estimating methodology, which was proposed in (Khan et al.,
2017), involving crowd sourcing approach as well as utilization of
conditional probability distribution of traffic states at adjacent junc-
tions (Khan et al., 2017).
Several other researches focused on proposing MCS based systems
for parking availability management. For example, ParkNet which was
proposed in (Mathuret al., 2010) to detect available parking spots in
cities using ultrasonic sensing devices installed on cars combined to
smart phones. ParkSense represents another smartphone based appli-
cation that detects the vacated parkingspots based on ubiquitous Wi-Fi
beacons (Nawaz, Efstratiou, & Mascolo, 2013). PE. Carnelli et al. in-
troduced a novel vehicle parking activity detection method, called
ParkUs to eventually reduce vacant car parking space search times.
ParkUs utilizes accelerometer and magnetometer sensors found in all
smartphones in order to detect parking activity within a city environ-
ment (Carnelli et al., 2017). In (Abe, Ogitsu, & Mizoguchi, 2016), a
novel automatic parking system was proposed where drivers cooperate
to indicate a parking space from the outside of a vehicle and assist the
vehicle in recognizing its environment so that the vehicle can park
semi-automatically. JG. Krieg et al. proposed SmartPark, a smartphone
based system that relies exclusively on the smartphone sensors to au-
tomatically point drivers to the nearest parking spot and enables
charging them exactly for the amount of time parked (Krieg et al.,
2016). S. Noor el al. introduced ParkBid, a crowd sourcing based
parking service for automobiles where the information of free parking is
circulated among the interested users following a bidding process
(Noor, Hasan, & Arora, 2017). In (Cherian et al., 2016), the authors
presented ParkGauge, a crowd sensing-based method to gauge the oc-
cupancy of parking garages by leveraging low-power sensors (e.g., ac-
celerometer and barometer) in the driver's smartphone. F. BOCK et al.
presented an evaluation of the suitability of a fleet of taxis as probe
vehicles for parking crowd-sensing. The authors exploited a dataset of
real-world trajectories collected from about 500 taxis over 3 weeks in
San Francisco (USA), to extract their movement patterns (Bock,
Attanasio, & Di Martino, 2017).
Ground transportation monitoring has also attracted some research
works. In (Kong et al., 2017), Y. Kong et al. proposed a system named
CrackDetector to detect cracks and estimate their types and size with
smart phone. F. Seraj et al. present a new algorithm for map matching
of crowd-sourced measurements for monitoring ground transportation
infrastructures using smartphones (Seraj, Meratnia, & Havinga, 2017).
355
 Table 1
A summary of main MCS applications for environmental monitoring.
Ref Usage Description Involvement of the users in the Application domain
sensing process
D.E. Boubiche et al.

(http://opensignal.com/.) Temperature monitoring A real-time temperature monitoring system to collect air temperatures based on an Android Medium Space weather
application for smartphones (Open Signal). monitoring
(Niforatos et al., 2014) Temperature monitoring Atmos: A hybrid crowdsourcing approach to weather estimation. The main idea is to leverage a High
crowd-sourcing network of mobile devices for the collection of in-situ weather related sensory data,
provided by available on-board sensors, along with human input, to generate highly localized
information about current and future weather conditions.
(Meier et al., 2017) Temperature monitoring An experimental study to examine the suitability of crowd sourced air temperature measurements Low
from citizen weather stations. A large number of measurement readings, from up to 1500 stations
with reference air temperature in Berlin and surroundings, have been compared to a period of
twelve months (Jan–Dec 2015).
(Elmore et al., 2014) Winter precipitation mPING: An efficient crowd sensing app for smartphones which allows citizen scientists to provide High
monitoring observations about the winter precipitation type at the surface at least equivalent in quality to
human-augmented Automated Surface Observing System (ASOS) observations.
(Mass, Clifford, Madaus & Luke, 2014) Pressure monitoring PressureNet: MCS based application that collects atmospheric pressure measurements from its users, Medium
with the aim of using this data to help understand the atmosphere and better predict the weather.
Ref Usage Description Involvement of the users in the Application domain
sensing process

(Honicky, Brewer, Paulos, & White, 2008)
(Antonić et al., 2014)
Air pollution N-Smarts: pollution project, which uses sensors attached to GPS enabled smart phones to gather Medium Air pollution monitoring
data, in order to help to better understand how urban air pollution impacts both individuals and
communities.
Air quality monitoring Energy efficient urban crowd sensing application for air quality monitoring based on the use of Low
wearable sensors and mobile devices. The application also provides a personalized real-time alert
mechanism to mobile users.

356
(Mun, Reddy, Shilton, & et al, 2009) Carbon dioxide PEIR: a participatory sensing application that uses location data sampled from everyday mobile Medium
monitoring phones to calculate personalized estimates of environmental impact and exposure.
(Goldman et al., 2009) Air pollution A mobile crowd application which uses data gathering software run on residents' mobile phones to Low
gather the information of on-the-road diesel trucks to study community exposure to urban air
pollution.
(Predic, Yan, Eberle, Dragan Stojanovic, & Pollution exposure ExposureSense: a hybrid mobile participatory sensing infrastructure that integrates the WSN and Low
Aberer, 2013) estimation MCS paradigms. ExposureSense is able to monitor people's daily activities as well to compute a
reasonable estimation of pollution exposure in their daily life.
(Zheng, Liu, & Hsieh, 2013) Air quality monitoring UAir: inferred fine-grained air quality information of a city based on heterogeneous crowd sourced Low
data, including the air quality, traffic flow, human mobility, structure of road networks and points
of interest (POIs).
(Eisenman et al., 2009) CO2 pollution monitoring BikeNet: an extensible mobile sensing system which leverages personal, bicycle, and environmental Medium
sensing using dynamic role-assigned bike area networking to measure and report the CO2 pollution
level.
(Sivaraman et al., 2013) Air pollution monitoring HazeWatch: A Participatory low-cost participatory system for urban air pollution monitoring in Medium
Sydney. Several low-cost mobile sensor units attached to vehicles are used to measure the air
pollution concentrations, and the users' mobile phones are leveraged to tag and upload the data in
real time.
Ref Usage Description Involvement of the users in the Application domain
sensing process

(continued on next page)

Computers in Human Behavior 101 (2019) 352–370
D.E. Boubiche et al. Computers in Human Behavior 101 (2019) 352–370

4.3. Social MCS applications

An account of human powered applications of MCS applications

Application domain

revealed multiple aspects e.g., individuals share sensed information

Noise pollution

amongst themselves involving recommendations and opinions, life ex-

monitoring

periences and service/activity suggestions. Several social MCS (Zhang,

Choo, Sangaiah, & Chen, 2018) based applications have been proposed
for venue recommendation where the historical location trajectories
recorded by mobile devices are leveraged. For example, Y. Zheng et al.
Involvement of the users in the

proposed a location-history-based recommender system, which uses a

particular individual's visit on a geospatial location as their implicit
ratings on the location, to predict a particular user's interest in an un-
visited location (Zheng et al., 2011). Another place recommendation
system was proposed in (Ye et al., 2011) to provide a point-of-interests
sensing process

(POI) recommendation service for the rapid growing location-based

social networks. In (Wirz et al., 2011a), the authors proposed a new
Medium

system for real-time detection and recommendation of thermal spots.

High

High

High

High

High

High
Low

Low

The main idea is to simultaneously aggregate flight information from

NoiseSPY: turns the mobile phone into a low-cost data logger for monitoring environmental noise. It

NoiseBattle & NoiseQuest: MCS application for gathering noise pollution data. The applications are

NoizCrowd: allowing to crowd source noise levels in a given region and to generate noise models by

where all the generated information from the voluntary participants on a city scale are aggregated

many paraglider pilots using their location-aware mobile phones. Y.

designed following gamification techniques to encourage users to participate using their personal

Ear-Phone: a context-aware noise mapping using Smart phones.Ear-Phone investigates the use of

recovering the noise map from incomplete and random samples obtained by crowdsourcing data
NoiseMap: smart phone application based on participatory sensing leading to accurate, real-time

MCS-based system allowing users to gather noise measurements in order to perform large-scale,
loudness of their environment. The measured data are anonymised and sent to a central server

Chon et al. presented CrowdSense@Place (CSP), a framework that ex-

SoundOfTheCity: a smart phone application that allows the users to continuously measure the

MOOL: a platform that adopts the MCS paradigm for involving secondary schools' students in
allows users to explore a city area while collaboratively visualizing noise levels in real-time.

ploits a previously untapped resource opportunistically captured

different interpolation and regularization methods to address the fundamental problem of
NoiseTube: enables citizens to measure their personal exposure to noise in their everyday

using state-of-the-art noise propagation models and array data management techniques.

images and audio clips from smartphones to link place visits with place
categories (e.g., store, restaurant) (Chon et al., 2012). CheckInside was
introduced by M. Elhamsharyet al. to provide a fine-grained indoor
location-based social network. CheckInside exploits the crowd-sensed
information collected from users' mobile devices during the check-in
low-cost and sufficiently accurate urban noise monitoring campaigns.
environment by using GPS-equipped mobile phones as noise sensors.

operation and knowledge extracted from current LBSNs (location-based

social networks) to associate a place with its name and semantic fin-
gerprint (Elhamshary & Youssef, 2014). Also, Z. Yu et al. proposed a
didactic experimentation about acoustics and noise sensing.

travel package recommendation system to help users make travel plans

by leveraging mobile crowd sourced data (Yu et al., 2014). Another
and mapped to a meaningful noise visualization map.

venue recommendation application, called SIPF (Social-aware Inter-

esting Place Finding), was proposed in (Huang & Wang, 2015) where
user's travel experience and social relationship are mainly exploited. An
enhanced version of SIPF was also proposed in (Huang, Dong & Mann,
2017), namely SIPF+. Recently, C Huang el al. presented an un-
supervised approach to accurately discover interesting places in a city
from location based social sensing applications which collect users'
observations of the physical world (Chao & Wang, 2016). CrowdTravel
was also recently proposed to provide travel assistance for tourists by
crowd intelligence mining (Guo et al., 2017a).
Another example of social application in MCS is the DietSense
smartphones.
noise maps.
Description

collection.

system (Reddy et al., 2007), where users take pictures of what they eat
and share it within a community to compare their eating habits. A
classical usage example for this is for a community of diabetics to
lookout what other diabetics eat and control their diet or deliver re-
Personal noise monitoring

commendations to others. CrowdMonitor represents also a distinctive

Noise visualization map

Urban noise monitoring

Noise levels monitoring

Real-time noise maps

example of social MCS application. CrowdMonitor was designed to

Reginal noise levels

create coordination mechanisms for interacting and collaborating with

Noise pollution

Noise pollution

Noise pollution

the public during emergencies (Ludwig et al., 2015a).

monitoring

monitoring

monitoring

monitoring
Usage

4.4. Behavioral applications

Another outcome of human-powered sensing applications categor-

ized under behavioral domain revealed various interesting aspects. The
authors in (Fujisaka, Lee, & Sumiya, 2010) created social designs using
(Ruge, Altakrouri, & Schrader, 2013)

client histories from micro-bloggers and propose models based on

conglomeration and scattering. These social approaches are used to find
geographic characteristics accumulated from Twitter to study crowd's
(Maisonneuve et al., 2009)

(Zappatore et al., 2016b)

(Wisniewski et al., 2013)

(Zappatore et al., 2016a)

(Schweizer et al., 2011)

behavior. In (Wakamiya, Lee, & Sumiya, 2012), the authors in-

Table 1 (continued)

vestigated crowd's urban lifestyles and its behavior using Twitter, and
(Martí et al., 2012)

(Rana et al., 2015)

propose a system model for a large-scale urban analytics with the lo-
(Kanjo, 2010)

cation-based social network represented by crowd's temporal beha-

vioral patterns. Wirz et al. in (Wirz et al., 2011b) investigated how real-
time participatory MCS can be exploited to infer collective behavior
Ref

patterns and to conclude about community intelligence. In this

357

Table 2
A summary of main MCS applications for infrastructure monitoring.
Ref Usage
(Graunke, 2001) Traffic analysis
(Mohan et al., 2008) Traffic congestion
(Wen et al., 2008) Traffic congestion
(Liu et al., 2009) Traffic analysis
(Thiagarajan et al., 2010) Transit tracking
(Lin et al., 2010) Comfort measuring
(Zhou et al., 2012) Traffic predition
(Hu et al., 2013) Traffic analysis
(Farkas et al., 2014) Traffic analysis
(Bröring et al., 2015) Traffic analysis
358
(Wan et al., 2016) Traffic predition
(Yan et al., 2017) Traffic congestion
(Kaur et al., 2017) Traffic analysis
(Khan et al., 2017) Traffic predition
(Mathuret al., 2010) Parking statistics
(Nawaz et al., 2013) Vacated parkingspots
(Carnelli et al., 2017) Vacated parkingspots
(Abe et al., 2016) Automatic parking
(Krieg et al., 2016) Nearest parking spot
(Noor et al., 2017) Parking reservation
(Cherian et al., 2016) Vacated parkingspots
(Bock et al., 2017) Vacated parkingspots
(Kong et al., 2017) Cracks detection
(Seraj et al., 2017) Transportation infrastructures
monitoring
D.E. Boubiche et al.
Description Involvement of the users in the sensing Application domain
process
Predictive traffic modeling Medium Traffic management
Nericell: rich monitoring of road and traffic conditions using mobile smartphones Medium
Operational analysis on Beijing road network during the Olympic Games Medium
Real time monitoring of urban mobility patterns Medium
Cooperative transit tracking using smart-phones High
A comfort measuring system for public transportation systems using participatory High
phone sensing
Predicting bus arrival time with mobile phone based participatory sensing High
SmartRoad: a mobile phone based crowd-sourced road sensing system Medium
TrafficInfo: Participatory sensing based real-time public transport information High
service
enviroCar: A Citizen Science Platform for Analyzing and Mapping Crowd‐Sourced High
Car Sensor Data
Mobile crowd sensing for traffic prediction in internet of vehicles Medium
Cloud-Assisted Mobile Crowd Sensing for Traffic Congestion Control Low
Traffic state detection using smartphone based acoustic sensing Low
RoadCrowd: An approach to road traffic forecasting at junctions using crowd- Medium
sourcing and Bayesian model
Parknet: detecting available parking spots in cities using ultrasonic sensing devices Low Parking management
Parksense: A smartphone based sensing system for on-street parking Low
ParkUs: A Novel Vehicle Parking Detection System Low
Parking Position Detection with Human Cooperation for Automatic Parkin High
SmartPark: a smartphone based system that points drivers to the nearest parking spot Medium
ParkBid: An Incentive Based Crowdsourced Bidding Service for Parking Reservation High
ParkGauge: Gauging the Occupancy of Parking Garages with Crowdsensed Parking Low
Characteristics
Crowd-Sensing of Parking Availability with Taxis High
CrackDetector: Detecting Type and Size of Road Crack with the Smartphone. Medium Ground transportation monitoring
An aggregation and visualization technique for crowd-sourced continuous Medium
monitoring of transport infrastructures
Computers in Human Behavior 101 (2019) 352–370
D.E. Boubiche et al. Computers in Human Behavior 101 (2019) 352–370

research, the authors reveal that real-time detection of collective be-

Venue recommendation
havior patterns that emerge among the pilots leads to discover regions

Medical & Emergency

with ideal thermal characteristics, which can help pilots to extend their
Application domain

Travel assistance
flight time and to fly longer distances.
In (Guo et al., 2015b), Bin et al. address human urban mobility/
behavior patterns by analyzing the check-in histories of a huge set of
LBSN (location-based social organize) clients. An application named
EmotionSense was created by capturing the user's emotive, behavioral,
and social signals, based on which one's real-time nature can be dis-
tinguished. Jia et al. recognized the MCS potential with behaviors
Involvement of the users in the sensing

perspective in (Jia, Jnxin, & Yang, 2015). In this study, authors re-
ference various applications and systems, such as DietSense and Bi-
keNet for healthcare, CenceMe and Co-evolution model for behavior
and relationship discovery, PIER for calculating personalized environ-
mental impact and exposure. The authors in (Ludwig et al., 2015b)
developed applications of mobile crowd sensing include emergency
behavior pattern and its usefulness.
An application of MCS paradigm in supporting efficient, safe and
Medium

Medium

Medium

Medium
Medium
Medium

Medium
process

green mobility in urban environments with behavior perspectives was

High
High

High

High

presented in (Stojanovic et al., Purves). They develop CitySensing fra-

mework demonstrating the viability of a common crowd-sourcing
Image browsing, processing, and clustering for participatory sensing: lessons from a DietSense
Real-time detection and recommendation of thermal spots by sensing collective behaviors in

CrowdMonitor: Mobile crowd sensing for assessing physical and digital activities of citizens

platform applied to various urban mobility domains. In their MCS re-

CrowdSense@Place: Automatically characterizing places with opportunistic crowdsensing

search, authors reveal to crowd source diverse information in beha-

CrowdTravel: Unsupervised interesting places discovery in location-based social sensing
Exploiting geographical influence for collaborative point-of-interest recommendation

vioral domains that are relevant to urban life and mobility (traffic, air
quality and citizens' everyday activities). For Crowd-traffic sensing,
system detect and report traffic events, road infrastructure, driver be-
SIPF+: social-aware interesting place finding in social sensing applications
Recommending friends and locations based on individual location history

havior and activities, and accidental events. For Air quality sensing, the
ExposureSense application sensed air-quality parameters (like tem-
Recommending travel packages based on mobile crowdsourced data

perature, concentration of different particles or gasses in the air). Urban

CheckInside: A fine-grained indoor location-based social network.

authorities can use such crowd intelligence to improve urban mobility

by detecting mobility behaviors and patterns of citizen's/tourists
movement. For citizen's behavior pattern sensed using UrbanSense
application served as a mobile diary service that detects and stores
information about users' environment and mobile interaction by col-
lecting relevant information.
SIPF: Social-aware Interesting Place Finding

The authors in (Anjomshoa & Kantarci, 2018) simulated the socia-

bility behavior of participants in MCS scenarios using sociability pro-
files for battery power saving. In (Guo et al., 2014b; Zhang, Xiong,
Wang, & Chen, 2014), authors develop predictive models for the dis-
covery of events, communities and knowledge to study behavior pat-
terns. In (Anjomshoa et al., 2016), authors preciously identify beha-
during emergencies
using smartphones

vioral patterns of smartphone/tablet users in which social networking

applications were the target platforms to mine behavioral patterns.
Description

paragliding

MCS was employed in (Liu et al., 2018b) to investigate the behavioral

prototype

pattern with IoT. In (Atzori, Iera, & Morabito, 2010), Atzori et al.
concluded that IoT would have a high impact on potential users' be-
havior. In (Gorlatova et al., 2014), Gorlatova et al. analyze a motion
Point-of-interests recommendation
Point-of-interests recommendation

dataset and develop an energy allocation algorithm with accessible IoT

Travel packages recommendation
Thermal spots recommendation

node solution in order to characterize the energy availability related to

particular human behaviors. In (Liu et al., 2009), Liu et al. monitor the
Travel recommendation

mobility patterns of citizens and visualize data to show development

Places characterizing

Places characterizing

trends in a city's economy and infrastructure specific to resident

A summary of main social applications for MCS.

Emergency control

transport behavior.
Place finding
Place finding

Diet control

In (Adeel, Yang, & McCann, 2014), Adeel el al. address how to

provide a cost-effective networking service for real-time and delay-
Usage

tolerant applications in Mobile Urban Sensing System for mobility/

behavior patterns. In (Rabbi, Ali, Choudhury, & Berke, 2011), authors
present a MCS system for measuring mental well-being from behavioral
(Elhamshary & Youssef, 2014)

indicators in natural everyday settings as primary healthcare e.g.

(Huang & Wang, 2015)

weight loss. In (Dong, Hoover, Scisco, & Muth, 2012), authors present a
(Ludwig et al., 2015a)
(Huang et al., 2017)

(Reddy et al., 2007)

(Zheng et al., 2011)

(Wirz et al., 2011a)

MCS system for measuring intake via automated tracking of wrist mo-
(Guo et al., 2017a)
(Chon et al., 2012)

(Yu et al., 2014)

(Ye et al., 2011)

tion in healthcare.

5. MCS architecture
Table 3

Ref

Depending on the application, different MCS architectures have

359
D.E. Boubiche et al. Computers in Human Behavior 101 (2019) 352–370

Table 4
A summary of some behavior applications for MCS.
Ref Usage Description Application domain

(Fujisaka et al., 2010)
(Wakamiya et al., 2012)
(Wirz et al., 2011b)
(Guo et al., 2015b)
(Jia et al., 2015)
(Ludwig et al., 2015b)
(Stojanovic et al., Purves)
(Anjomshoa & Kantarci, 2018)
(Guo et al., 2014b; Zhang et al.,
2014a)
(Anjomshoa et al., 2016)
(Liu et al., 2018b)
(Atzori et al., 2010)
(Gorlatova et al., 2014)
(Liu et al., 2009)
(Adeel et al., 2014)
(Rabbi et al., 2011)
(Dong et al., 2012)
Geolocation characterizing
Point-of-interests recommendation
Thermal spots recommendation
Emotion characterizing
Point-of-interests recommendation
Point-of-interests recommendation
Traffic, air pollution and citizen
behavior characteristics
Social characteristics
Point-of-interests recommendation
Point-of-interests recommendation
Point-of-interests recommendation
Point-of-interests recommendation
Energy saving
Point-of-interests recommendation
Networking Services
Basic healthcare with weight loss
Measuring intake in healthcare
Develop social design based on client histories from micro- Geographical characteristics
blogger via Twitter.
Exploiting urban lifestyle and behavior pattern using Location based social network
Twitter.
Crowds infer collective behavior pattern to conclude Travelling assistance
community intelligence.
Exploit behavior pattern by analyzing check-in history Urban mobility
location based social client
Address potential of MCS references such as DietSence, Healthcare, behavior and relation
BikeNet, CenseMe, Co-evolution and PIER applications discovery, environment
Address emergency behavior pattern and its usefulness Emergency behavior
Develop CitySensing framework based on Crowd- traffic, Urban lifestyle and mobility
ExposeSence and UrbanSence applications
Simulate social behavior of participants in MCS using their Power saving for mobile battery
social profile
Develop predictive model for discovery of events, Community behavior
communities and knowledge.
Mining behavior pattern of smartphone users with social Social network
networking application
Investigate behavior pattern with MCS in IoT Behavior pattern with IoT
Concluded that IoT has high impact on potential user Social network with IoT
behavior
Develop energy saving algorithm for IoT node selection to Energy saving with IoT
characterize energy availability.
Monitor citizen mobility pattern to show the development Urban development
in city economics and infrastructure.
Address how to provide network services for Mobile Urban Mobile urbanization
Sensing System.
Develop MCS system for mental well-being for behavior Healthcare
pattern
MCS system to measure intake via automated tracking of Healthcare
wrist motion in healthcare

been adopted [ (Ganti et al., 2011), (Surowiecki, 2004), (Lane et al.,
2010), (Trossen & Pavel, 2007), (Mun et al., 2009) and (Hu et al.,
2014)]. However, to make it more general, a typical MCS application
has three basic levels, according to the provided services: 1) Data
sensing and storage; 2) data processing and analytics; 3) application
and services. Fig. 3 presents the three basic levels of the typical MCS
architecture model.
5.1. Data sensing and storage level
This level refers to the sensing and the storage infrastructures and
the involved sensing components. Diverse crowd sensing sources are
considered, including mobile sensor devices and Internet applications.
For this reason, this level provides a standard data collection and
common interface to enable crowd sensing from heterogeneous sources.
The storage of the sensing data is another service supported by this
level where the sensed data are archived and formatted according to a
cohesive presentation. Indeed, due to the huge volume, which might be
collected through different sources (e.g. collecting weather information
in large city), it will be difficult and inappropriate to use classical data
management tools. In addition to the data collection and storage, this
level provides privacy protection service for sensing and shared per-
sonal data.
5.2. Processing and analytics level
This level is responsible of data analysis, learning and mining. The
collected sensory data are processed to extract and leverage the desired
information using data mining and machine learning techniques. Also,
extracted data will be filtered to eliminate redundancy and ensure ac-
curacy. Indeed, the processing tasks might take place locally on the
mobile device or on the mobile cloud. Furthermore a hybrid processing
architecture can be applied with some partitioning between the mobile
phone and the cloud.
This level covers all kinds of MCS applications and services. Also, it
defines the user's interfaces and display the services ensuring the pre-
sentation of the crowd processed results in a comprehensive format to
the final user. In other words, this level is considered as an interface
between the backend crowd sensing platform and the users.
MCS architecture can be also divided, according to the interacting
entities, into three main components: the server, the mobile providers
and the clients. The server component represents a centralized platform
which schedules and attributes sensing tasks to be executed on the in-
dividual's mobile platforms. Also, the server accepts task requests from
clients and then forward them to mobile provider according to the es-
tablished schedule. The mobile providers represent mobile platforms
which participate to the crowd sensing process to provide the sensing
data requested by the clients. The clients designate the end users (or-
ganization or group of users) which leverage the mobile crowd sensing
system to collect and sense data from different mobile devices (e.g.,
smart phones and PDAs). The communication between all these com-
ponents occurs through the network. Fig. 4, illustrates the MCS archi-
tecture according to the involved entities (see Fig. 5).
Based on the data collecting and dissemination approach, data pre-
processing and task allocation model, MCS architectures can be also
classified into centralized or decentralized. Indeed, in the centralized
MCS architecture, a central unit takes in charge the management of the
data sensing and the sending process. In the other side, the decen-
tralized MCS architecture enables users to fully control their own per-
sonal data (e.g., decide with whom they are going to share them and
when), with the risk however of acquiring potentially low quality data.
6. Challenges and solutions
The unique characteristics of MCS have introduced numerous ad-
vantages and benefits as well as several challenges and research

360
D.E. Boubiche et al. Computers in Human Behavior 101 (2019) 352–370

Table 5
Summary of solutions for human MCS involvement.
Proposed solutions Challenges handled

Solution description Rewarding strategie

SenseMart: Sensing Data Market that stimulates user participation (Chou et al., 2007) Reputation rankings Incentive mechanism
RADP: Reverse Auction based Dynamic Price mechanism (Lee & Baik, 2010) Financial rewards
Rewarding models for data acquisition and distributed computing (Duan et al., 2012) Financial rewards
Privacy based incentive schemes (Li & Cao, 2016) Financial rewards
APISENSE: an incentive platform for scientists, and participants (Haderer et al., 2013) Participant rankings, involvement badges
BBS: A behavior based incentive mechanism with budget constraints (Sun & Ma, 2013) Financial rewards
iMac: incentive mechanism to prevent overpayment (Feng et al., 2013) Financial rewards
SPREAD: incentive assignment mechanism Financial rewards
To Acquire Better Representative Samples (Jaimes et al., 2014)
Privacy-respecting auctions as incentive mechanisms in the mobile crowd sensing (Dimitriou & Financial rewards
Krontiris, 2015)
A Quality Based Incentive Mechanism for Crowd sensing (Peng et al., 2015) Financial rewards
INCEPTION: Incentivizing Privacy-Preserving Data Aggregation for MCS (Jin et al., 2016) Financial rewards
An incentive game based evolutionary model for crowd sensing networks (Liu et al., 2016) Gamification
Source-centric and user-centric based incentive mechanism for MCS (Yang et al., 2016) Financial rewards
BidGuard:A Framework for Privacy-Preserving Crowd sensing Incentive Mechanisms (Lin et al., Ranking and social recognition
2016)
Promoting cooperation by the social incentive mechanism in the mobile crowd sensing (Yang et al., Ranking and social recognition
2017)
A three-layer incentive framework for the mobile crowd sensing (Chen et al., 2017a) Ranking and social recognition
A Truthful Incentive Mechanism for Online Recruitment (Chen et al., 2017b) Gamification
Sustainable Incentives designs for Mobile Crowd sensing (Luo et al., 2017) Bargaining games, reputation rankings and
financial rewards
TaskMe: dynamic and quality-enhanced incentive mechanism (Guo et al., 2017b) Social recognition and financial rewards
ESWM: Expected Social Welfare Maximizing (Back, Choi, & Chen, 2017) Financial rewards
Practical Incentive Mechanisms for IoT-Based Financial rewards
Mobile Crowdsensing Systems (Duan, Li, & Cai, 2017)
HySense: A Hybrid Mobile CrowdSensing Framework for Sensing Opportunities Compensation Financial rewards
under Dynamic Coverage Constraint (Han et al., 2017)
RIT: Robust Incentive Tree Design for Mobile Financial rewards
Crowdsensing (Zhang, Xue, Yu, & et al, 2017a, 2017b)
Geo-QTI: A quality aware truthful incentive mechanism for cyber–physical enabled Geographic Gamification and financial rewards
crowdsensing (Dai et al., 2017)
A QoS-Aware Online Incentive Mechanism for Mobile Crowd Sensing (Cai, Zhu, & Yu, 2017) Financial rewards
Privacy-Aware Incentive Mechanism for Mobile Crowd Sensing (Koh et al., 2017) Financial rewards
TAFA: an auction-based incentive mechanism for crowdsourcing (Zhang et al., 2017a, 2017b) Financial rewards
FIMI: A Constant Frugal Incentive Mechanism for Time Window Coverage in Mobile Crowdsensing Financial rewards
(Xu, Shen, Zhu, & et al, 2017)
TSIA: Time-sensitive incentive-aware mechanism for mobile opportunistic crowdsensing data Bargaining games
collection (Zhan et al., 2017)
An Incentive Mechanism for Crowdsensing Markets with Multiple Crowdsourcers (Chakeri & Jaimes, Gamification and financial rewards
2017a)
PIE: A personalized incentive for location-aware mobile crowd sensing (Wu et al., 2017) Financial rewards
A Double Auction Mechanism for Mobile Crowd Sensing with Data Reuse (Zhang et al., 2017a, Financial rewards
2017b)
An iterative incentive mechanism design for crowd sensing using best response dynamics (Chakeri & Gamification and financial rewards
Jaimes, 2017b)
Solution description Selection/Filter criteria Task allocation
Dynamic data driven crowd sensing task assignment (Pournajaf, Xiong, & Sunderam, 2014) Region coverage
Location Privacy-Preserving Task Allocation for Mobile User privacy
Crowdsensing (Wang et al., 2017)
Recruitment framework for participatory sensing data collections (Reddy, Estrin, & Srivastava, Region coverage and user behavior
2010)
CrowdRecruiter: participant selection framework for MCS (Zhang et al., 2014b) Sensing cost and region coverage
Optimal Allocation of Location Dependent Tasks in Crowdsensing (He et al., 2014) Processing and sensing cost
CCS-TA: Quality-guaranteed online task allocation in compressive crowd sensing (Wang et al., 2015) Region coverage
Multi-task assignment for crowd sensing (Xiao et al., 2015) Processing and sensing cost
iCrowd: Near-optimal task allocation for piggyback crowd sensing (Xiong et al., 2016) Sensing cost and region coverage
Budget-feasible online incentive mechanisms for crowdsourcing tasks truthfully (Zhao, Li, & Ma, Sensing cost and user privacy
2016)
Real-time task assignment in hyper local spatial crowdsourcing under budget constraints (To et al., Region coverage (spatiotemporal)
2016)
Activecrowd: A framework for optimized multitask allocation in MCS (Guo et al., 2017c) Region coverage (spatiotemporal)
Distributed auction schemes for task allocation and scheduling (Duan Tian & Yan, 2017) Sensing cost and region coverage
Mew: A Plug-n-Play Framework for Task Allocation in Mobile Crowdsensing (Bajaj & Pushpendra, User privacy
2017)
Solution description Grouping strategie Participant interactions
Virtual team formation approach (Vukovic, 2009) Request based grouping
Community-aware smartphone sensing systems (Lane, 2012) User skills and attributes based grouping
GroupMe: Supporting Group Formation with Mobile Sensing (Guo et al., 2012) Social graph mining
MobiGroup: Mobil group activity organization in MCS (Guo et al., 2016a) Activity based grouping

361
D.E. Boubiche et al.
Fig. 3. A typical MCS architecture.
opportunities. Indeed, implementing MCS in the real world is complex
and can bring several problems related to security and the privacy,
human involvement and collaboration, data accuracy, delivery, re-
dundancy and quality. In this section, we provide a general overview of
the main MCS challenges and prominent relevant solutions from the
literature.
6.1. Human sensing involvement
Integrating human volunteers in the MCS process (e.g., data sensing
and gathering, sharing information) represents a real challenge. Indeed,
incentive mechanisms are highly needed to incite the human partici-
pation and keep the users engaged. Also, managing the human inter-
actions, the event coverage, the group sensing and the task allocation
are important points that need to be addressed.
• Incentive mechanisms:
Most of MCS applications require voluntary participation. However,
contributing to the crowd sensing process is a resource consuming task
(e.g. battery and computing power). Also, in most cases, volunteers
have to share private information about themselves (i.e., data with tag
location) which might expose them to several security threats.
Therefore, MCS would be not attractive for users unless they receive
sufficient recompenses to pay back the potential resource consumption
and privacy breach. Indeed, without efficient incentives, the users may
not participate in such applications. It is a challenge to attract appro-
priate users and keep them involved while sensing the measurements
on the smart phone. Different incentive strategies can be applied to
362
Computers in Human Behavior 101 (2019) 352–370
attracting users. Most of these strategies are based on financial rewards
or ranking and social recognition. Also, incentive strategies can be
based on attention-grabbing task (gamification), enjoyment and al-
truism services.
Several research works, which investigate the incentive issues in
MCS, have been proposed in the literature. In the following, we give a
short survey of these incentive solutions based on the adopted re-
warding strategie.
6.1.1. Ranking and social reputation based incentive mechanisms
SenseMart (Chou, Bulusu, & Kanhere, 2007), represents one of the
first research works that addressed the use of reputation based incentive
mechanisms to facilitate the exchange of sensing data among users (as
in a marketplace). APISENSE represents another ranking based in-
centive platform that encourages scientists and participants to con-
tribute to the sensing experiments. Scientist incentives include the
modular integration of a wide variety of features they can compose
upon their need, while the participant incentives deal with privacy and
rewarding mechanisms that can be offered in order to catalyze the
collection of datasets (Haderer, Rouvoy, & Seinturier, 2013). A general
privacy-preserving framework for incentivizing the crowd sensing,
named BidGuard, was proposed in (Lin et al., 2016). The authors con-
sidered both the bid privacy of the smartphone users and the social cost
to design an efficient incentive mechanism which achieves better
computational efficiency, individual rationality, truthfulness, differ-
ential privacy and approximate social cost minimization. Recently, G.
Yang et al. introduced a novel approach, called the social incentive
mechanism, which, surprisingly, incentivizes the social friends of the
participants who perform the sensing tasks. The basic idea is to leverage
D.E. Boubiche et al.
Fig. 4. MCS architecture according to the involved entities.
Fig. 5. MCS architecture according to the centralized/decentralized control.
the social ties among the participants to promote global cooperation
(Yang et al., 2017). Similarly to (Yang et al., 2017), Chen, Y et al.
proposed to incorporate sensing platform and social network applica-
tions, which already have large user bases to build a three-layer net-
work model. Thus, the sensing platform can be publicized promptly in
large scale, and provide long-term guarantees of data sources (Chen
et al., 2017a). Luo, T. et al. provided an exposition of design principles
of six incentive mechanisms, drawing special attention to the sustain-
ability issue. The authors covered three primary classes of incentive
mechanisms: auctions, lotteries, and trust and reputation systems, as
well as three other frameworks of promising potential: bargaining
games, contract theory, and market-driven mechanisms (Luo et al.,
2017). Guo, B. et al. proposed a novel MCS incentive mechanism called
TaskMe. The main idea is to leverage an LBSN (location-based social
network)-powered model for dynamic budgeting and proper worker
selection, and to improve the sensing quality by the use of a combi-
nation of multi-facet quality measurements and a multi-payment-en-
hanced reverse auction scheme (Guo et al., 2017b).
6.1.2. Financial based incetive mechanisms
In order to stimulate the user participation, J.S Lee et al. designed a
novel Reverse Auction based Dynamic Price (RADP) incentive me-
chanism, where the users can sell their sensing data to a service pro-
vider with users' claimed bid prices (Lee & Baik, 2010). A comparative
study was presented in (Duan et al., 2012) to analyze different incentive
mechanisms for motivation and collaboration of smartphone users. The
authors also introduced two rewarding models for the data acquisition
and the distributed computing. Q Li et al. addressed the security aspect
in the incentive mechanism and proposed two privacy-aware incentive
schemes. The proposed schemes allow each mobile user to earn credits
by contributing data without leaking information, and ensure at the
same time that the dishonest users cannot abuse the system to earn
unlimited amount of credits (Li & Cao, 2016). J Sun et al. investigated
the negligence of the data quality and privacy during the incentive
363
Computers in Human Behavior 101 (2019) 352–370
mechanism design and proposed a behavior based incentive mechanism
which maximizes the high quality of sensing data under the budget
constraint (Sun & Ma, 2013). iMac was introduced in (Feng, Zhu, & Ni,
2013) for incentive crowd sensing where the users are stimulated to
truthfully disclose their real costs, preventing thus user misreporting
cost (different from the real cost) and overpayment. L.G. Jaimes et al.
proposed SPREAD, an incentive crowd sensing mechanism designed to
obtain the lowest cost samples that are best distributed to cover the area
of interest within a fixed budget (Jaimes, Vergara-Laurens, & Chakeri,
2014). In (Dimitriou & Krontiris, 2015), the authors presented a reverse
auction mechanism as an efficient way to offer incentives to the users
by allowing them to determine their own price for the data they pro-
vide, but also as a way to motivate them to submit the data in better
quality (Dimitriou & Krontiris, 2015). In (Peng, Wu, & Chen, 2015), the
authors incorporated the consideration of data quality into the design of
incentive mechanism for crowd sensing, and propose to pay the parti-
cipants as how well they do, to motivate the rational participants to
perform data sensing efficiently (Peng et al., 2015). INCEPTION, a
novel MCS system framework that integrates an incentive, a data ag-
gregation, and a data perturbation mechanism was proposed in (Jin
et al., 2016). The main idea is to select the workers who are more likely
to provide reliable data, and compensates their costs for both sensing
and privacy leakage. Yang, D. et al. designed incentive mechanisms
based on two system models: the crowd source-centric model where the
crowd source provides a reward shared by participating users, and the
user-centric model where the users have more control over the payment
they will receive (Yang et al., 2016).
6.1.3. Gamification based incetive mechanisms
X. Liu et al. proposed an incentive game based on an evolutionary
model for the crowd sensing networks. Through the evolutionary game
model, the manager can select an optimal price to facilitate the system
to reach equilibrium state quickly, and get the number of participants
involved in the game (Liu et al., 2016). Chen, X. et al. designed an
efficient and truthful incentive mechanism to encourage the users to
participate. To address the challenge, the authors proposed a novel
truthful online auction mechanism that can efficiently learn to make
irreversible online decisions on winner selections for new MCS systems
without requiring previous knowledge of users (Chen et al., 2017b)
• Task allocation:
To ensure efficient crowd sensing, task assignment approaches are
applied to select a set of qualified individuals to perform the sensing
tasks. Indeed, participant selection introduces more challenges in the
MCS and can have impact on the efficiency and quality of the sensing
results. Several criteria are used to filter the inappropriate participants
such as the available device competencies, region coverage (e.g. se-
lecting sensing devices situated in the desired monitoring region),
sensing cost and participant privacy. Commonly, a task assignment
framework includes three main entities: 1. Participants which represent
individuals who use a sensor to obtain or measure the required data
about a subject of interest. 2. Applications or end users, which are the
entities that request data through tasks and then utilize the information
acquired by the participants. 3. Tasking entities which are responsible
for distribution of tasks to participants who meet the requirements of
applications (Pournajaf et al., 2014).
Based on the user selection criteria we summarize most important
proposed task allocation approaches as follow:
6.1.4. Spatiotemporal based task allocation solutions
Reddy S. et al. discussed the development of a recruitment frame-
work to enable the organizers to identify the well-suited participants for
data collections based on geographic and temporal availability as well
as participation habits (Reddy et al., 2010). Pournajaf, L. et al. proposed
a novel model for the spatial task assignment in the MCS that uses a
D.E. Boubiche et al.
dynamic and adaptive data driven scheme to assign moving partici-
pants with uncertain trajectories to sensing tasks, in a near-optimal
manner (Pournajaf et al., 2014). Wang, L. et al. proposed a novel fra-
mework called CCS-TA, combining the state-of-the-art compressive
sensing, Bayesian inference, and active learning techniques, to dyna-
mically select a minimum number of sub-areas for sensing task allo-
cation in each sensing cycle, while deducing the missing data of non
allocated sub-areas under probabilistic data accuracy guarantee (Wang
et al., 2015). Xiong, H.et al. introduced iCrowd, a generic MCS task
allocation framework operating with the energy-efficient Piggyback
Crowd sensing task model to optimize the MCS task allocation with
different incentives and k-depth coverage objectives/constraints (Xiong
et al., 2016). To, H. et al. introduced a task allocation framework in
which only the workers who are already within the spatiotemporal
vicinity of a task are eligible candidates to report data, e.g., the pre-
cipitation level at their area and time. In this setting, a subset of can-
didate workers whose size is constrained by a predefined budget can be
activated to perform tasks (To et al., 2016). Recently, Guo, B. et al.
studied the worker selection problem under multi-task MCS environ-
ments and have identified two common situations: intentional-move-
ment-based selection for time-sensitive tasks and unintentional-move-
ment-based selection for delay-tolerant tasks. The optimal goal for the
two situations varies and the authors proposed two greedy-enhanced
genetic algorithms GGA-I, GGA-U to address them (Guo et al., 2017c).
Also, Duan, Z. et al. studied the joint problem of sensing task assign-
ment and scheduling while considering partial fulfillment, attribute
diversity, and price diversity. Furthermore, the authors proposed two
distributed auction schemes, cost-preferred auction scheme (CPAS) and
time schedule-preferred auction scheme (TPAS) which optimize the
sensing task's allocation efficiency, mobile user's working time utiliza-
tion and utility, and truthfulness (Duan et al., 2017).
6.1.5. Sensing const and device competencies
Crowd Recruiter represents another proposed task allocation fra-
mework which operates on top of energy-efficient Piggyback Crowd
sensing (PCS) task model and minimizes the incentive payments by
selecting a small number of participants while still satisfying prob-
abilistic coverage constraint (Zhang et al., 2014b). He, S. et al. defined
the task allocation problem as NP-hard and designed a pricing me-
chanism based on bargaining theory, in which the price of each task is
determined by the performing cost and the market demand (He et al.,
2014). In (Xiao et al., 2015), the authors proposed two task allocation
algorithms: an oFfline Task Assignment (FTA) algorithm and an oNline
Task Assignment (NTA) algorithm. Both of the algorithms adopt a
greedy task assignment strategy to achieve the minimum average of the
task carrying out. Zhao, D et al. designed two online task assignment
mechanisms, OMZ and OMG, satisfying six desirable properties: com-
putational efficiency, individual rationality, budget feasibility, truth-
fulness, consumer sovereignty and constant competitiveness (Zhao
et al., 2016).
6.1.6. User privacy
There are some preliminary studies on task allocation design for
MCS based on user privacy. WANG Leye et al. proposed a location
privacy-preserving task allocation framework with geoobfuscation to
protect the users' locations during the task assignments (Wang et al.,
2017). Bajaj, G et al. proposed Mew, which offers plug-n-play func-
tionality for implementing the custom task allocation algorithms to
allow developers to reach out to the required set of participants while
solving privacy challenge associated with MCS application in a black-
box manner (Bajaj & Pushpendra, 2017).
• Participant interactions:
Managing participant interactions represents an important step for
the optimal crowd sensing process. Indeed, grouping users, facilitates
364
Computers in Human Behavior 101 (2019) 352–370
the interaction among them and might enhance the data quality in
MCS. However, managing user's interactions is a complex task that was
neglected in most existing crowd sensing systems. Only few research
works addressed this issue (Vukovic, 2009). One of first research works
that addressed the group formation challenge in MCS was presented in
(Vukovic, 2009). The authors proposed a virtual team formation ap-
proach that employs an expertise matching mechanism and discovered
social relationships. Lane, N. D. et al. also pointed out the importance of
participants interactions and group formation in designing better equip
sensing systems. The authors also exposed the need of automated
methods to identify which user interactions and ties are the most salient
to the operation of the sensing system (Lane, 2012). Guo, B. et al.
proposed another participants grouping approach called GroupMe,
which facilitates the activity group initiation based on mobile sensing
and social graph mining (Guo et al., 2012). A group-aware mobile
crowd sensing system called MobiGroup was proposed in (Guo et al.,
2016a), which supports group activity organization in real world set-
tings. The authors present a formal concept model to characterize group
activities and classify them into four organizational stages. Also an
intelligent approach to support group activity preparation was in-
troduced.
6.2. Data quality issues
Ensuring high data quality without redundancy and inconsistency
represents a big challenge for MCS systems. Indeed, data quality, in
terms of accuracy, latency, completeness and sparsity, is highly im-
pacted by MCS characteristics such as participant mobility, device
owners' preferences and the type of sensor data each one can produce
and the variation in the communication channels.
• Redundant and low quality data:
MCS involves a large amount of individual volunteer in the sensing
process which may lead to data redundancy. Therefore, it is imperative
to efficiently select data from multiple available sources. Indeed, sev-
eral sensing sources may collect the same data with different quality
levels due to the variation of the sensing conditions. Thus, applying
data selection and filtering approach is needed to reject low-quality or
irrelevant data. To exclude the negative impact of abnormal or low-
quality data in crowd sensing, Ding et al. introduced a data cleansing-
based robust spectrum sensing algorithm (Ding et al., 2014). Messaoud
R. M. et al. proposed another selection scheme for participatory sensing
tasks, named QEMSS, to select, among all participants in the sensing
campaigns, the subset of users who maximizes the quality of informa-
tion of non-redundant information while minimizing the overall energy
consumption (Messaoud & Ghamri-Doudane, 2015). Marjanović, M.
et al. presented a framework for Green MCS (G-MCS) which utilizes a
quality-driven sensor management function to continuously select the
k-best sensors for a predefined sensing task. In particular, it removes
redundant sensor activity while satisfies sensing coverage requirements
and sensing quality (Marjanović et al., 2016). PicPick, is another data
selection approach which aims to eliminate data redundancy and re-
duce the network overhead. PicPick first presents a multifaceted task
model that allows sensing task specification for variable data. A pyr-
amid tree (PTree) method is further proposed to select an optimal set of
data from data streams based on multi-dimensional constraints (Guo
et al., 2016b). Recently, Meng, C. et al. developed an integrated fra-
mework to estimate the true values of entities from redundant and
sparse data in crowd sensing applications (Meng et al., 2016). The
authors proposed an effective algorithm to infer the “missing” ob-
servations for each entity, and aggregate both user-contributed and
deducted observations to discover the true values of entities.
• Data accuracy and completeness:
D.E. Boubiche et al.
The incompleteness and correctness of sensory data pose several
challenging issues for efficient MCS process. Indeed, only in an ideal
environment, participants provide accurate and complete sensing
readings. Kurasawa et al. pointed out that the data collected by crowd
usually include considerable missing values in practical crowd-sensing
systems. They proposed a method to estimate missing values using a
recursive regression model (Kurasawa et al., 2014). several research
progresses in missing data inference algorithms, such as compressive
sensing, could facilitate MCS to achieve high inferred data quality
(Quer et al., 2012), (Zhu et al., 2013), (Kong et al., 2013), (Xu et al.,
2015a). To address the problem of low quality, misleading and in-
accurate data, Yang, H. et al. proposed to use a reputation management
approach to classify the gathered data and provide useful information
for campaign organizers and data analysts to facilitate their decisions
(Yang, Zhang, & Roe, 2011). Recently, Cheng L. et al. proposed Deco, a
general framework to detect false values for crowd-sensing in the pre-
sence of missing data. By applying a tailored spatio-temporal com-
pressive sensing technique, Deco is able to accurately detect the false
data and estimate both false and missing values for data correction
(Cheng et al., 2017). Santosh K. et al. proposed an intelligent decision
computing based paradigm for crowd monitoring in the smart city. In
the intelligent computing based framework, the optimization algorithm
is applied to compute the feature of crowd motion and measure the
correlation between agents based motion model and the crowd data
using extended Kalman filtering approach and KL-divergence technique
(Kumar, Datta, Singh, & Sangaiah, 2018).
6.3. Security and privacy
Alongside the data quality and human involvement concerns, se-
curity is rated as a top challenge issue. As.
MCS applications involve data collection across a large number of
human participants, several security threats may occur targeting mainly
data privacy and integrity. Therefore, it is imperative to ensure data
integrity and privacy without restraining MCS applications.
6.3.1. Threat model and security attacks in MCS
Degrading the MCS service level, crashing the system or stealing
privacy data are the main objectives of an attacker. Also, MCS can be
targeted by an internal attacker which aims to extract secrets or obtain
rewards with less sensing efforts. On the other hand, more complex
attackers can apply machine learning techniques to perform attacks
with selected policy and generate then dangerous threats which can
overcome the MCS security defenses. In the following, we briefly review
some significant types of threats in MCS.
6.3.1.1. Faked sensing attacks. Under-sensing or faked sensing reports
can be occasionally submitted by selfish mobile users to save sensing
efforts and protect privacy.
6.3.1.2. Malwares. Resource constrained mobile users are particularly
vulnerable to malwares. Indeed, malwares such as spy tools, viruses or
worms seriously threaten MCS with the network performance
degradation, economic loss and privacy leakage.
6.3.1.3. Spoofing attacks. In this kind of attacks, the identity of a
legitimate mobile device or an access point can be used by the
attacker to illegally access the MCS system. Also, other attacks, such
as man-in-the-middle and DoS attacks, can be performed based on the
spoofed identity.
6.3.1.4. Sybil attacks. A large number of different user identities can be
used by a Sybil user to send sensing reports and influence the sensing
result made by the majority-based MCS server.
6.3.1.5. Privacy leakage. User privacy must be protected by MCS
365
Computers in Human Behavior 101 (2019) 352–370
servers particularly the user locations and the personal information.
Indeed, privacy information can be stolen by rogue users from
transmission processes or cloud resources which may discourage the
other mobile users from participating to the MCS tasks.
6.3.1.6. DoS attack. This attack targets mainly the MCS services in
order to reduce the performing level and exhaust the MCS server
resources.
6.3.1.7. Jamming. The main objective is to interrupt the ongoing
transmission of the sensing data by injecting faked or replay signals.
To overcome the anti-jamming technics, the attacker can use multiple
jamming power and frequency according to the transmission status.
6.3.2. MCS security issues and countermeasures
• Data privacy:
The sensing data collected by mobile devices may inadvertently
reveal personal information about the mobile users, including identity,
location, habit, health status, daily route, and political affiliation.
Furthermore, the more tasks mobile users engaged in and the richer
data they contribute to, the higher possibility that their sensitive in-
formation is disclosed (Ni, Zhang, Lin, & et al, 2017). Several research
works addressed the privacy problem in MCS applications. A popular
approach for preserving the privacy of the data is that of anonymization
which removes any identifying information from the sensor data before
sharing it with a third party (Ganti et al., 2011). However, privacy
attack can be lanced based on other sensitive information like GPS
position. Cryptographic solutions represent another alternative to pre-
vent privacy attacks, though, this kind of security approaches are
compute intensive and unsuitable for the resource limited MCS devices.
To effectively preserve the privacy, not only methodology efforts, but
also systematic studies are needed. In other words, a privacy-aware
architecture should be provided to support the development of MCS
applications (Surowiecki, 2004). Cornelius, C. et al. proposed one of the
first privacy-aware architectures for MCS named AnonySense. Anony-
Sense allows applications to submit sensing tasks that will be dis-
tributed across anonymous participating mobile devices based on a
stringent threat model that adopts minimal trust assumptions
(Cornelius et al., 2008). Konidala, D. M. et al. focused on both au-
thentication and privacy aspects and proposed an anonymous authen-
tication of visitor's protocol that protects the privacy of visitors even
while collecting their details, preferences and location coordinates
(Konidala et al., 2013). Krontiris, I. et al. suggested a mobile node
discovering mechanism that corresponds to range queries about a
geographical area. This mechanism enables mobile users to protect
their location privacy, according to their own preferences and enforce
access policies to their own data (Krontiris & Dimitriou, 2013).
SPPEAR, a comprehensive, secure and privacy-preserving architecture,
was proposed in (Gisdakis, Giannetsos, & Papadimitratos, 2014) to
address security and privacy concerns in participatory sensing systems.
SPPEAR guarantees the user non-identifiability and offers strong
privacy protection, limits participation to legitimate users in a fully
accountable manner and efficiently shuns out offending users without,
necessarily, revealing their identity (Gisdakis et al., 2014). Another
privacy solution was introduced in (Krontiris & Dimitriou, 2015) were
the authors proposed a mobile sensing platform that protects the
anonymity of both data requesters and producers, while enabling at the
same time the incorporation of trust frameworks, incentive mechanisms
and privacy-respecting reputation schemes. Miao, C. et al. proposed a
novel cloud-enabled privacy-preserving truth discovery (PPTD) frame-
work for crowd sensing systems, which can achieve the protection of
not only users' sensory data but also their reliability scores derived by
the truth discovery approaches (Miao et al., 2015). Wu, X. et al. pre-
sented a privacy preserving RSS map generation scheme for crowd
D.E. Boubiche et al.
sensing networks called PRESM. To protect the privacy of user traces,
the authors exploit the compressive sensing technique to sample and
compress RSS (Received Signal Strength) values along each road seg-
ment, which removes the temporal and concrete location information of
each participant (Wu et al., 2015). PPCS, a novel Privacy Preserving
Compressive Sensing scheme, was introduced by Kong, L. et al. to
preserve the user privacy. The main idea is to perturb a trajectory with
several other trajectories while maintaining the homomorphic obfus-
cation property for compressive sensing. Under PPCS, adversaries can
only capture the encrypted data (Kong et al., 2015). Shin, M. et al.
proposed and evaluate a novel spatiotemporal blurring mechanism
based on tessellation and clustering to protect users' privacy while re-
porting context. The proposed technique employs a notion of prob-
abilistic k-anonymity; it allows users to perform local blurring of re-
ports efficiently without an online anonymization server before the data
are sent to the system (Shin et al., 2015). The privacy of trajectory
information was also addressed in (Zhang et al., 2016a) where the
authors proposed a participant coordination framework, which allows
the system server to provide optimal QoI (Quality of Information) for
sensing tasks without knowing the trajectories of participants. A com-
prehensive security and privacy-preserving architecture was proposed
in (Gisdakis et al., 2014). The main idea is to extend the SPPEAR pro-
tocol to optimize the user privacy while supporting user incentive
mechanisms. INCEPTION, represents a novel MCS system framework
that integrates an incentive, a data aggregation, and a data perturbation
mechanism. Specifically, its incentive mechanism selects workers who
are more likely to provide reliable data, and compensates their costs for
both sensing and privacy leakage. Data perturbation mechanism en-
sures satisfactory protection for workers' privacy and desirable accu-
racy for the final perturbed results (Jin et al., 2016). Recently, Ni, J.
et al. proposed a privacy-preserving mobile crowd sensing framework
(PPMC) for location-based applications to balance the tradeoff between
privacy preservation and task allocation. In PPMC, the authors devel-
oped a matrix-based location matching mechanism for the service
provider to achieve location-based task allocation without disclosing
the location of mobile users and the sensing area of tasks (Ni et al.,
2017). Nasir R. et al. proposed a privacy-preserving image retrieval for
mobile devices with deep features on the cloud. The authors proposed
to represent images with hash codes, which is a compressed re-
presentation of deep convolutional features using deep auto-encoder on
the cloud. To ensure user's privacy, the image is first encrypted using a
light-weight encryption algorithm on mobile device prior to offloading
it to the cloud for features extraction (Rahim, Ahmad, Muhammad,
Sangaiah, & Baik, 2018).
• Data integrity:
Data integrity represents another security aspect which must be
addressed to ensure the integrity of sensed data from individual's smart
phones. Only few research works focused on the data integrity problem
in MCS applications and the field remains undertreated. Dua A. et al.
proposed and implemented a trusted platform module (TPM), which is
a micro-controller embedded within each mobile device, to attest the
integrity of the sensor readings. However, TPM chips are not yet widely
adopted in mobile devices (Dua et al., 2009). Xu, J. et al. investigated
truthful incentive mechanisms for time window dependent tasks in the
mobile crowd sensing based on strong requirement of data integrity (Xu
et al., 2015b). Multidisciplinary approaches to achieve efficient and
trustworthy eHealth monitoring systems was introduced in (Sawand
et al., 2014). The authors also proposed a light-weight security me-
chanism for eHealth monitoring, to ensure authenticity, confidentiality,
and integrity of medical data collected from patients. To address the
privacy and data integrity problems, Zhang, L. et al. proposed an effi-
cient data aggregation approach by which an untrusted aggregator in
mobile sensing can collect the statistics over the data contributed by
multiple mobile users, while supporting privacy preservation of each
366
Computers in Human Behavior 101 (2019) 352–370
user and data integrity verification. In this approach, information
hiding and homomorphic encryption are applied to guarantee the data
privacy of mobile users (Zhang et al., 2016b). Yadav, K. et al. In-
troduced Human Sensors, an architecture of crowd sensing testbed for
capturing and processing events affecting citizens in cities in India. The
authors checked the integrity and the authenticity of a report by
comparing it to other reports collected from independent participants
(Yadav et al., 2013). Gilbert, P. et al. proposed a trustworthiness model
which focuses on data integrity (for data coming from automatic
readings from devices), data correctness, and quality. User's contribu-
tions are compared to those provided by local authoritative data
sources, certified by the data provenance micro service (Gilbert et al.,
2010). A novel data aggregation protocol in the crowd sensing system
was introduced in (Xu, Fu & Yang 2017), to protect the data integrity
and achieve authentication and traceability by leveraging the aggregate
signature. Also, Chen, J. et al. proposed a novel private data aggrega-
tion scheme to address secure data-integrity verification problem while
taking into account the security vulnerability of limited data range in
MCS applications (Chen et al., 2017c).
• Data authentication:
Authenticating the identity of the MCS participants represents an
important security issue which contributes to prevent several security
threats such as spoofing and Sybil attacks. The identity authentication
involves confirming the legitimate identity of the users to obtain au-
thorization for accessing the MCS system. Konidala, D. M. et al. ad-
dressed the authentication issue in MCS and introduced an anonymous
authentication algorithm to authenticate the identity of the visitor
without revealing his/her identity. The fundamental idea is to interact
with the visitor using a pseudonym instead of his/her true identity
(Vukovic, 2009). Based on a similar concept, Ma, P. et al. proposed a
pseudonym based anonymous identity authentication mechanism for
MCS. The introduced solution contributes to the privacy protection of
the authentication process by the means of pseudonym. Also, the au-
thors proposed to combine the Public Key Infrastructure (PKI) and the
Combined Public Key (CPK) technology for key and certificate man-
agement to solve the problem of the large-scale key management (Ma,
Tao, & Wu, 2017). An enhanced secure certificateless privacy-preser-
ving verifiable data authentication scheme for mobile crowd sensing,
named EPDA, was proposed by Liu, J.et al. The proposed scheme pro-
vides unconditional anonymous data authentication service for mobile
crowd sensing, by deploying an improved certificateless ring signature
as the cryptogram essential, in which the big sensing data should be
signed by one of the legitimate members in a specific group and could
be verified without exposing the actual identity of the participant (Liu
et al., 2017). Another anonymous authentication scheme was presented
In (Dinkar & Jain, 2016). It provides anonymous authentication to the
users who share their important data to any third part. An original
authentication strategy was introduced recently by Shi, C. et al. where
the unique physical features of WiFi signals in the daily activities of
mobile users is captured to identify the mobile users. More specifically,
the system extracts 6 time domain features, such as maximum,
minimum and skewness, and 3 frequency domain features, including
spectrogram energy, percentile frequency component and spectrogram
energy difference, from both of the amplitude and the phase of the
channel responses of the WiFi signals (Shi et al., 2017).
• Data reliability:
Ensuring the data reliability of the participants' sensed data in
crowd sensing is a crucial security issue, since malicious users can
submit false data to earn money without executing the actual task.
Therefore, there is a need for mechanisms to efficiently validate the
collected data. However, the sensing measurements are highly depen-
dent on context; therefore, it is challenging to validate each and every
D.E. Boubiche et al.
sensed data point of each participant. To handle this issue, several re-
search works are proposed to validate the location associated with the
sensed data point in order to achieve a certain degree of reliability on
the sensed data. However, this kind of solution is confronted to a major
challenge which is the location validation of the data points in a scal-
able and cost-effective way without help from the wireless carrier.
Some traditional approaches have been proposed to ensure reliability
on participants' location data such as tasks duplication among multiple
participants or using Trusted Platform Modules (TPM) (Trusted plat-
form module.). However, these solutions are usually not recommended
due to the introduced cost (installing TPM modules on every smart
phone) and feasibility. Another solution to verify the reliability on
participants' location data is to exploit the location verification me-
chanisms (Capkun, ˇ Cagalj, & Srivastava, 2006; Capkun & Hubaux,
2005). However, more support infrastructures are required and user
phones may suffer from significant overhead.
Instead of using data location to guarantee the reliability of the
sensed data, another approache mainly, based on data mining techni-
ques, aims to discover the ground truth by analyzing the noisy and
possibly conflicting MCS data (sometimes taking workers into account
as well). Wang et al. used the Expectation-Maximization (EM) algo-
rithm to find the maximum likelihood estimate (MLE) of the probability
that a given binary MCS measurement is true (Wang, Kaplan, Le, &
Abdelzaher, 2012). Davami and Sukthankar combined multiple trust-
based data fusion techniques using AdaBoost, a machine learning al-
gorithm, to predict whether a parking lot is occupied, based on the
crowdsourced user reports (Davami & Sukthankar, 2015). Gisdakis
et al. proposed SHIELD to perform outlier detection in the presence of
compromised sensing devices, but it requires a large amount of data to
train the complex machine learning model (Gisdakis, Giannetsos, &
Papadimitratos, 2015).
Recently, an original new solution for data reliability was proposed
by Li, D. et al. The authors introduced a new cross validation approach
into MCS. This approach does not mean comparing a piece of data
(contributed by a worker as in the MCS context) against another piece
of data (contributed by another worker), which is similar to the ap-
proach used in the machine learning and data mining literature.
Instead, the new proposed cross validation approach means subjecting
the crowd sensed data to a group of crowdworkers who did not con-
tribute to the original MCS data to seek their verification, for the pur-
pose of reshaping or “grooming” the original sensed data toward the
ground truth (Li et al., 2018).
7. Conclusion
Mobile crowd sensing represents one of the most impactful tech-
nologies which will transform radically several sectors of our economy,
including environmental monitoring and transportation, business,
healthcare and social networks. MCS exploits the sensor's smartphones
and the mobility of participants to sense their surroundings, instead of
deploying static and expensive sensor network. Moreover, existing
communication infrastructures (e.g., 3G, Wi-Fi etc.) are used by MCS
applications, to collect data from mobile phones scattered in the mon-
itoring field. MCS will bring promising opportunities, along with new
challenges. In this paper, we addressed the mobile crowd sensing
concept and presented its main features and advantages, architectures
and application domains. Also, we have identified the main MCS
challenges and gave an overview on the recent mobile crowd sensing
strategies and solutions.
Acknowledgement
This work is supported by the Deanship of Scientific Research, King
Saud University through Research Group No. RG-1439-036.
367
Computers in Human Behavior 101 (2019) 352–370
References
Abe, Y., Ogitsu, T., & Mizoguchi, H. (2016). Design and evaluation of parking position
detection with human cooperation for automatic parkin. Journal of the Institute of
Industrial Applications Engineers, 4(2), 100–105.
Adeel, U., Yang, S., & McCann, J. A. (2014). Self-optimizing citizen-centric mobile urban
sensing systems. Proceedings of the USENIX ICAC.
Anjomshoa, F., Catalfamo, M., Hecker, D., Helgeland, N., Rasch, A., Kantarci, B., et al.
(July 2016). Mobile behaviometric framework for sociability assessment and iden-
tification of smartphone users. Proceedings of the IEEE symposium on computers and
communications, messina, Italy (pp. 1084–1089). 27 June–1.
Anjomshoa, F., & Kantarci, B. (2018). SOBER-MCS: Sociability-Oriented and battery ef-
ficient recruitment for mobile crowd-sensing. MDPI Sensors, 18, 1593. https://doi.
org/10.3390/s18051593.
Antonić, A., Bilas, V., Marjanović, M., et al. (2014). Urban crowd sensing demonstrator:
Sense the zagreb air. Software, telecommunications and computer networks (SoftCOM),
2014 22nd international conference on (pp. 423–424). IEEE.
Atzori, L., Iera, A., & Morabito, G. (2010). The internet of things: A survey. Computer
Networks, 54(15), 2787–2850.
Back, D., Choi, B. J., & Chen, J. (2017). Small profits and quick returns: A practical
SocialWelfare maximizing incentive mechanism for deadline-sensitive tasks in crowd-
sourcing. arXiv preprint arXiv:1707.00018.
Bajaj, G., & Pushpendra, S. (2017). Mew: A plug-n-play framework for task allocation in
mobile crowdsensing. Proceedings of the first ACM workshop on mobile crowdsensing
systems and applications (pp. 19–24). ACM.
Bock, F., Attanasio, Y., & Di Martino, S. (2017). Spatio-temporal road coverage of probe
vehicles: A case study on crowd-sensing of parking availability with taxis.
International conference on geographic information science (pp. 165–184). Cham:
Springer.
Boulos, Kamel, M. N., Resch, B., Crowley, D. N., et al. (2011). Crowdsourcing, citizen
sensing and sensor web technologies for public and environmental health surveil-
lance and crisis management: Trends, OGC standards and application examples.
International Journal of Health Geographics, 10(1), 67.
Bröring, A., Remke, A., Stasch, C., et al. (2015). enviroCar: A citizen science platform for
analyzing and mapping crowd-sourced car sensor data. Transactions in GIS, 19(3),
362–376.
Cai, H., Zhu, Y., & Yu, J. (2017). A QoS-aware online incentive mechanism for mobile
crowd sensing. International conference on web information systems engineering (pp. 37–
45). Cham: Springer.
Capkun, S., & Hubaux, J. (2005). Secure positioning of wireless devices with application
to sensor networks. INFOCOM’05. 24th annual joint conference of the IEEE computer
and communications societies. Proceedings IEEE: Vol. 3, (pp. 1917–1928). IEEE.
Capkun, S., ˇ Cagalj, M., & Srivastava, M. (2006). Secure localization with hidden and
mobile base stations. proceedings of IEEE INFOCOM. Citeseer.
Carnelli, P. E., Yeh, J., Sooriyabandara, M., et al. (2017). ParkUs: A novel vehicle parking
detection system. AAAI (pp. 4650–4656). .
Chakeri, A., & Jaimes, L. (2017a). An incentive mechanism for crowdsensing markets
with multiple crowdsourcers. IEEE Internet of Things Journal, 5(2), 708–715.
Chakeri, A., & Jaimes, L. G. (2017b). An iterative incentive mechanism design for crowd
sensing using best response dynamics. Communications (ICC), 2017 IEEE international
conference on (pp. 1–7). IEEE.
Chao, H., & Wang, D. (2016). Unsupervised interesting places discovery in location-based
social sensing. Distributed computing in sensor systems (DCOSS), 2016 international
conference on (pp. 67–74). IEEE.
Chen, Y., Chen, H., Yang, S., et al. (2017). Jump-start crowdsensing: A three-layer in-
centive framework for mobile crowdsensing. Quality of service (IWQoS), 2017 IEEE/
ACM 25th international symposium on (pp. 1–6). IEEE.
Cheng, L., Niu, J., Kong, L., et al. (2017). Compressive sensing based data quality im-
provement for crowd-sensing applications. Journal of Network and Computer
Applications, 77, 123–134.
Chen, X., Liu, M., Zhou, Y., et al. (2017). A truthful incentive mechanism for online re-
cruitment in mobile crowd sensing system. Sensors, 17(1), 79.
Chen, J., Ma, H., et, & Zhao, D. (2017). Private data aggregation with integrity assurance
and fault tolerance for mobile crowd-sensing. Wireless Networks, 23(1), 131–144.
Cherian, J., Luo, J., Guo, H., et al. (2016). ParkGauge: Gauging the occupancy of parking
garages with crowdsensed parking characteristics. Mobile data management (MDM),
2016 17th IEEE international conference on (pp. 92–101). IEEE.
Chon, Y., Lane, N. D., LI, F., et al. (2012). Automatically characterizing places with op-
portunistic crowdsensing using smartphones. Proceedings of the 2012 ACM conference
on ubiquitous computing (pp. 481–490). ACM.
Chou, C. T., Bulusu, N., et, & Kanhere, S. (2007). Sensing data market. Proceedings of
poster papers of 3rd IEEE international conference on distributed computing in sensor
systems (pp. 13). .
Cornelius, C., Kapadia, A., Kotz, D., et al. (2008). Anonysense: Privacy-aware people-
centric sensing. Proceedings of the 6th international conference on Mobile systems, ap-
plications, and services (pp. 211–224). ACM.
Dai, W., Wang, Y., Jin, Q., et al. (2017). Geo-QTI: A quality aware truthful incentive me-
chanism for cyber–physical enabled geographic crowdsensing. Future Generation
Computer Systems.
Davami, E., & Sukthankar, G. (2015). Improving the performance of mobile phone
crowdsourcing applications. AAMAS (pp. 145–153). .
Dimitriou, T., & Krontiris, I. (2015). Privacy-respecting auctions as incentive mechanisms
in mobile crowd sensing. IFIP international conference on information security theory
and practice (pp. 20–35). Springer International Publishing.
Ding, G., Wang, J., Wu, Q., et al. (2014). Robust spectrum sensing with crowd sensors.
D.E. Boubiche et al.
IEEE Transactions on Communications, 62(9), 3129–3143.
Dinkar, K. P., & Jain, S. A. (2016). Blacklisting User In Mobile Crowd Sensing, 03(04),
2334–2338.
Dong, Y., Hoover, A., Scisco, J., & Muth, E. (2012). A new method for measuring meal
intake in humans via automated wrist motion tracking. Applied Psychophysiology and
Biofeedback, 37(3), 205–215.
Dua, A., Bulusu, N., Feng, W.-C., et al. (2009). Towards trustworthy participatory sensing.
Proceedings of the 4th USENIX conference on Hot topics in security (pp. 8). USENIX
Association 8.
Duan, L., Kubo, T., Sugiyama, K., et al. (2012). Incentive mechanisms for smartphone col-
laboration in data acquisition and distributed computing. INFOCOM, 2012 proceedings
IEEE. IEEE1701–1709.
Duan, Z., Li, W., & Cai, Z. (2017). Distributed auctions for task assignment and scheduling
in mobile crowdsensing systems. Distributed computing systems (ICDCS), 2017 IEEE
37th international conference on (pp. 635–644). IEEE.
Duan, Z., Tian, L., Yan, M., et al. (2017). Practical incentive mechanisms for IoT-based mobile
crowdsensing systems. IEEE Access.
Eisenman, S. B., Miluzzo, E., Lane, N. D., et al. (2009). BikeNet: A mobile sensing system
for cyclist experience mapping. ACM Transactions on Sensor Networks, 6(1), 6.
Elhamshary, M., & Youssef, M. (2014). CheckInside: A fine-grained indoor location-based
social network. Proceedings of the 2014 ACM international joint conference on pervasive
and ubiquitous computing (pp. 607–618). ACM.
Elmore, K. L., Flamig, Z. L., Lakshmanan, V., et al. (2014). mPING: Crowd-sourcing
weather reports for research. Bulletin of the American Meteorological Society, 95(9),
1335–1342.
Farkas, K., Nagy, A. Z., Tomás, T., et al. (2014). Participatory sensing based real-time
public transport information service. Pervasive computing and communications work-
shops (PERCOM workshops), 2014 IEEE international conference on (pp. 141–144).
IEEE.
Feng, Z., Zhu, Y., & Ni, L. M. (2013). imac: Strategy-proof incentive mechanism for mobile
crowdsourcing. International conference on wireless algorithms, systems, and applications
(pp. 337–350). Berlin, Heidelberg: Springer.
Fujisaka, T., Lee, R., & Sumiya, K. (2010). Discovery of user behavior patterns from geo-
tagged micro-blogs. In: Proc. Of the 4th Intl. Conf. On Uniquitous information man-
agement and communication.
Ganti, R. K., Ye, F. et al. Lei, H. (2011). Mobile crowdsensing: Current state and future
challenges. IEEE Communications Magazine, 49(11).
Gilbert, P., Cox, L. P., Jung, J., et al. (2010). Toward trustworthy mobile sensing.
Proceedings of the eleventh workshop on mobile computing systems & applications (pp. 31–
36). ACM.
Gisdakis, S., Giannetsos, T., et, & Papadimitratos, P. (2014). Sppear: Security & privacy-
preserving architecture for participatory-sensing applications. Proceedings of the 2014
ACM conference on Security and privacy in wireless & mobile networks (pp. 39–50). ACM.
Gisdakis, S., Giannetsos, T., & Papadimitratos, P. (2015). Shield: A data verification fra-
mework for participatory sensing systems. In ACM conference on security & privacy in
wireless and mobile networks (WiSec).
Goldman, J., Shilton, K., Burke, J., et al. (2009). Participatory sensing: A citizen-powered
approach to illuminating the patterns that shape our world. White Paper: Foresight &
Governance Project1–15.
Gorlatova, M., Sarik, J., Grebla, G., Cong, M., Kymissis, I., & Zussman, G. (2014). Movers
and shakers: Kinetic energy harvesting for the internet of things. Proceedings of the
SIGMETRICS. Austin (pp. 407–419). .
Graunke, G. L. (2001). No 6Using predictive traffic modeling, Vol. 256U.S. Patent 577, 3
juill.
Guo, B., Chen, H., Yu, Z., et al. (2016b). PicPick: A generic data selection framework for
mobile crowd photography. Personal and Ubiquitous Computing, 20(3), 325–335.
Guo, B., Chen, H., Yu, Z., et al. (2017b). TaskMe: Toward a dynamic and quality-en-
hanced incentive mechanism for mobile crowd sensing. International Journal of
Human-computer Studies, 102, 14–26.
Guo, T., Guo, B., Ouyang, Y., et al. (2017a). CrowdTravel: Scenic spot profiling by using
heterogeneous crowdsourced data. Journal of Ambient Intelligence and Humanized
Computing, 1–10.
Guo, B., He, H., Yu, Z., et al. (2012). GroupMe: Supporting group formation with mobile
sensing and social graph mining. International conference on mobile and ubiquitous
systems: Computing, networking, and services (pp. 200–211). Berlin, Heidelberg:
Springer.
Guo, B., Liu, Y., Wu, W., et al. (2017c). Activecrowd: A framework for optimized mul-
titask allocation in mobile crowdsensing systems. IEEE Transactions on Human-
Machine Systems, 47(3), 392–403.
Guo, B., Wang, Z., Yu, Z., et al. (2015a). Mobile crowd sensing and computing: The review
of an emerging human-powered sensing paradigm. ACM Computing Surveys, 48(1), 7.
Guo, B., Wan, Y., Yen, N., Huang, R., & Zhou, X. (2015b). Mobile crowd sensing and
computing: The review of an emerging human-powered sensing paradigm. ACM
Computing Surveys, 48(No. 1) Article 7, August.
Guo, B., Yu, Z., Chen, L., et al. (2016a). MobiGroup: Enabling lifecycle support to social
activity organization and suggestion with mobile crowd sensing. IEEE Transactions on
Human-Machine Systems, 46(3), 390–402.
Guo, B., Yu, Z., Zhou, X., et al. (2014a). From participatory sensing to mobile crowd
sensing. Pervasive computing and communications workshops (PERCOM workshops),
2014 IEEE international conference on (pp. 593–598). IEEE.
Guo, B., Yu, Z., Zhou, X., & Zhang, D. (2014b). From participatory sensing to mobile
crowd sensing. Proceedings of the IEEE 2014 international conference on pervasive
computing and communications workshops, Budapest, Hungary (pp. 593–598). 22–24.
Haderer, N., Rouvoy, R., & Seinturier, L. (2013). A preliminary investigation of user in-
centives to leverage crowdsensing activities. Pervasive computing and communications
workshops (PERCOM workshops), 2013 IEEE international conference on (pp. 199–204).
368
Computers in Human Behavior 101 (2019) 352–370
IEEE.
Han, G., Liu, L., Chan, S., et al. (2017). HySense: A hybrid mobile CrowdSensing fra-
mework for sensing opportunities compensation under dynamic coverage constraint.
IEEE Communications Magazine, 55(3), 93–99.
He, S., Shin, D.-H., Zhang, J., et al. (2014). Toward optimal allocation of location de-
pendent tasks in crowdsensing. INFOCOM, 2014 proceedings IEEE (pp. 745–753).
IEEE.
Honicky, R. J., Brewer, E. A., Paulos, E., & White, R. M. (2008). N-SMARTS: Networked
suite of mobile atmospheric real-time sensors, NSDR’08, august 18, 2008. Washington,
USA: Seattle.
Huang, C., Dong, W., & Mann, B. (2017). Towards social-aware interesting place finding
in social sensing applications. Knowledge-based Systems, 123, 31–40.
Huang, C., & Wang, D. (2015). On interesting place finding in social sensing: An emerging
smart city application paradigm. Smart city/SocialCom/SustainCom (SmartCity), 2015
IEEE international conference on (pp. 13–20). IEEE.
Hu, X., Li, X., Ngai, E., et al. (2014). Multidimensional context-aware social network
architecture for mobile crowdsensing. IEEE Communications Magazine, 52(6), 78–87.
Hu, S., Liu, H., Su, L., et al. (2013). SmartRoad: A mobile phone based crowd-sourced road
sensing system.
Jaimes, L. G., Vergara-Laurens, I., & Chakeri, A. (2014). SPREAD, a crowd sensing in-
centive mechanism to acquire better representative samples. Pervasive computing and
communications workshops (PERCOM workshops), 2014 IEEE international conference on
(pp. 92–97). IEEE.
Jia, X., Jnxin, & Yang, D. (SEP 2015). Incentive mechanisms for time window dependent
tasks in mobile crowdsensing. IEEE Transactions on Wireless Communications,
13(No. 9).
Jin, H., Su, L., Xiao, H., et al. (2016). Inception: Incentivizing privacy-preserving data
aggregation for mobile crowd sensing systems. MobiHoc (pp. 341–350). .
Kanjo, E. (2010). Noisespy: A real-time mobile phone platform for urban noise monitoring
and mapping. Mobile Networks and Applications, 15(4), 562–574.
Kaur, A., Sood, N., Aggarwal, N., et al. (2017). Traffic state detection using smartphone
based acoustic sensing. Journal of Intelligent and Fuzzy Systems, 1–8 no Preprint.
Khan, S. Z., Rahuman, W. M. A., Dey, S., et al. (2017). RoadCrowd: An approach to road
traffic forecasting at junctions using crowd-sourcing and Bayesian model. Research
and innovation in information systems (ICRIIS), 2017 international conference on (pp. 1–
6). IEEE.
Khan, W. Z., Yang, X., Aalsalem, M. Y., et al. (2013). Mobile phone sensing systems: A
survey. IEEE Communications Surveys & Tutorials, 15(1), 402–427.
Koh, J. Y., Peters, G. W., Leong, D., et al. (2017). Privacy-aware incentive mechanism for
mobile crowd sensing. Communications (ICC), 2017 IEEE international conference on
(pp. 1–6). IEEE.
Kong, L., He, L., Liu, X.-Y., et al. (2015). Privacy-preserving compressive sensing for
crowdsensing based trajectory recovery. Distributed computing systems (ICDCS), 2015
IEEE 35th international conference on (pp. 31–40). IEEE.
Kong, L., Xia, M., Liu, X.-Y., et al. (2013). Data loss and reconstruction in sensor networks.
INFOCOM, 2013 proceedings IEEE (pp. 1654–1662). IEEE.
Kong, Y., Yu, Z., Chen, H., et al. (2017). Detecting type and size of road crack with the
smartphone. Computational science and engineering (CSE) and embedded and ubiquitous
computing (EUC), 2017 IEEE international conference on (pp. 572–579). IEEE.
Konidala, D. M., Deng, R. H., Li, Y., et al. (2013). Anonymous authentication of visitors
for mobile crowd sensing at amusement parks. International conference on information
security practice and experience (pp. 174–188). Berlin, Heidelberg: Springer.
Krieg, J.-G., Jakllari, G., Toma, H., et al. (2016). Unlocking the smartphone's senses for
smart city parking. Communications (ICC), 2016 IEEE international conference on (pp.
1–7). IEEE.
Krontiris, I., & Dimitriou, T. (2013). Privacy-respecting discovery of data providers in
crowd-sensing applications. Distributed computing in sensor systems (DCOSS), 2013
IEEE international conference on (pp. 249–257). IEEE.
Krontiris, I., & Dimitriou, T. (2015). A platform for privacy protection of data requesters
and data providers in mobile sensing. Computer Communications, 65, 43–54.
Kumar, S., Datta, D., Singh, S. K., & Sangaiah, A. K. (2018). An intelligent decision
computing paradigm for crowd monitoring in the smart city. Journal of Parallel and
Distributed Computing, 118, 344–358.
Kurasawa, H., Sato, H., Yamamoto, A., et al. (2014). Missing sensor value estimation
method for participatory sensing environment. Pervasive computing and communica-
tions (PerCom), 2014 IEEE international conference on (pp. 103–111). IEEE.
Lane, N. D. (2012). Community-aware smartphone sensing systems. IEEE Internet
Computing, 16(3), 60–64.
Lane, N. D., Miluzzo, E., Lu, H., et al. (2010). A survey of mobile phone sensing. IEEE
Communications Magazine, 48(9).
Lee, J.-S., & Baik, H. (2010). Sell your experiences: A market mechanism based incentive
for participatory sensing. Pervasive computing and communications (PerCom), 2010
IEEE international conference on (pp. 60–68). IEEE.
Li, Q., & Cao, G. (2016). Providing privacy-aware incentives in mobile sensing systems.
IEEE Transactions on Mobile Computing, 15(6), 1485–1498.
Lin, C.-Y., Chen, L.-J., Chen, Y.-Y., et al. (2010). A comfort measuring system for public
transportation systems using participatory phone sensing. ACM PhoneSense.
Lin, J., Yang, D., Li, M., et al. (2016). BidGuard: A framework for privacy-preserving
crowdsensing incentive mechanisms. Communications and network security (CNS),
2016 IEEE conference on (pp. 145–153). IEEE.
Li, D., Qiu, L., Liu, J., et al. (2018). Analysis of behavioral economics in crowdsensing: A
loss aversion cooperation model. Scientific Programming, 2018.
Liu, L., Biderman, A., & Ratti, C. (2009). Urban mobility landscape: Real time monitoring
of urban mobility patterns. Proceedings of the 11th international conference on compu-
ters in urban planning and urban management (pp. 1–16). .
Liu, J., Cai, F., Wu, L., et al. (2017). EPDA: Enhancing privacy-preserving data
D.E. Boubiche et al.
authentication for mobile crowd sensing. GLOBECOM 2017-2017 IEEE global com-
munications conference (pp. 1–6). IEEE.
Liu, G., Liu, S., Muhammad, K., Sangaiah, A. K., & Doctor, F. (2018). Object tracking in
vary lighting conditions for fog based intelligent surveillance of public spaces. IEEE Access.
Liu, X., Ota, K., Liu, A., et al. (2016). An incentive game based evolutionary model for
crowd sensing networks. Peer-to-Peer Networking and Applications, 9(4), 692–711.
Liu, J., Shen, H., Narman, H. S., Chung, W., & Zongfang, L. (June 2018). A survey of
mobile crowdsensing techniques: A critical component for the internet of things. ACM
Trans. Cyber-Phys. Syst. 2(3), 26 Article 18.
Ludwig, T., Reuter, C., Siebigteroth, T., et al. (2015a). CrowdMonitor: Mobile crowd
sensing for assessing physical and digital activities of citizens during emergencies.
Proceedings of the 33rd annual ACM conference on human factors in computing systems
(pp. 4083–4092). ACM.
Ludwig, T., Siebigteroth, T., & Crowd, V. P. (2015b). Monitor: Monitoring physical and
digital activities of citizens during emergencies. Social Influence, 88(52), 421–428.
Luo, T. T., Kanhere, S. S., Huang, J., et al. (2017). Sustainable incentives for mobile
crowdsensing. arXiv preprint arXiv:1701.00248.
Maisonneuve, N., Stevens, M., Niessen, M. E., et al. (2009). NoiseTube: Measuring and
mapping noise pollution with mobile phones. Information technologies in environmental
engineering215–228.
Marjanović, M., Skorin-Kapov, L., Pripužić, K., et al. (2016). Energy-aware and quality-
driven sensor management for green mobile crowd sensing. Journal of Network and
Computer Applications, 59, 95–108.
Martí, I. G., Rodríguez, L. E., Benedito, M., et al. (2012). Mobile application for noise
pollution monitoring through gamification techniques. International conference on
entertainment computing (pp. 562–571). Berlin, Heidelberg: Springer.
Mass, Clifford, F., Madaus, & Luke, E. (2014). Surface pressure observations from
smartphones: A potential revolution for high-resolution weather prediction? Bulletin
of the American Meteorological Society, 95(9), 1343–1349.
Ma, P., Tao, D., & Wu, T. (2017). A pseudonym based anonymous identity authentication
mechanism for mobile crowd sensing. Big data computing and communications
(BIGCOM), 2017 3rd international conference on (pp. 10–14). IEEE.
Mathur, S., et al. (2010). Parknet: Drive-by sensing of road-side parking statistics. In: In
Proc. of ACM MobiSys, 123–136.
Ma, H., Zhao, D., & Yuan, P. (2014). Opportunities in mobile crowd sensing. IEEE
Communications Magazine, 52(8), 29–35.
Mehmood, Y., Ahmad, F., Yaqoob, I., Adnane, A., Imran, M., & Guizani, S. (Sept, 2017).
Internet-of-Things based smart cities: Recent advances and challenges. IEEE
Communications Magazine, 55(9), 16–24.
Mehmood, Y., Ahmad, F., Yaqoob, I., et al. (2017). Internet-of-things-based smart cities:
Recent advances and challenges. IEEE Communications Magazine, 55(9), 16–24.
Meier, F., Fenner, D., Grassmann, T., et al. (2017). Crowdsourcing air temperature from
citizen weather stations for urban climate research. Urban Climate, 19, 170–191.
Meng, C., Xiao, H., Su, L., et al. (2016). Tackling the redundancy and sparsity in crowd
sensing applications. SenSys, 150–163.
Messaoud, R. B., et, & Ghamri-Doudane, Y. (2015). Qemss: A selection scheme for par-
ticipatory sensing tasks. Protocol engineering (ICPE) and international conference on new
technologies of distributed systems (NTDS), 2015 international conference on (pp. 1–6).
IEEE.
Miao, C., Jiang, W., Su, L., et al. (2015). Cloud-enabled privacy-preserving truth dis-
covery in crowd sensing systems. Proceedings of the 13th ACM conference on embedded
networked sensor systems (pp. 183–196). ACM.
Mohan, P., Padmanabhan, V. N., et, & Ramjee, R. (2008). Nericell: Rich monitoring of
road and traffic conditions using mobile smartphones. Proceedings of the 6th ACM
conference on Embedded network sensor systems (pp. 323–336). ACM.
Mun, M., Reddy, S., Shilton, K., et al. (2009). PEIR, the personal environmental impact
report, as a platform for participatory sensing systems research. Proceedings of the 7th
international conference on Mobile systems, applications, and services (pp. 55–68). ACM.
National Research Council of the National Academies (2012). Urban meteorology:
Forecasting, monitoring, and meeting users' need. Washington, U.S: Natl. Acad. Press.
Nawaz, S., Efstratiou, C., & Mascolo, C. (2013). Parksense: A smartphone based sensing
system for on-street parking. Proceedings of the 19th annual international conference on
Mobile computing & networking (pp. 75–86). ACM.
Niforatos, E., Vourvopoulos, A., Langheinrich, M., et al. (2014). Atmos: A hybrid
crowdsourcing approach to weather estimation. Proceedings of the 2014 ACM inter-
national joint conference on pervasive and ubiquitous computing: Adjunct publication (pp.
135–138). ACM.
Ni, J., Zhang, K., Lin, X., et al. (2017). Privacy-preserving mobile crowdsensing for lo-
cated-based applications. Communications (ICC), 2017 IEEE international conference on
(pp. 1–6). IEEE.
Noor, S., Hasan, R., & Arora, A. (2017). ParkBid: An incentive based crowdsourced bid-
ding service for parking reservation. Services computing (SCC), 2017 IEEE international
conference on (pp. 60–67). IEEE.
Peng, D., Wu, F., et, & Chen, G. (2015). Pay as how well you do: A quality based incentive
mechanism for crowdsensing. Proceedings of the 16th ACM international symposium on
mobile Ad Hoc networking and computing (pp. 177–186). ACM.
Pournajaf, L., Xiong, L., & Sunderam, V. (2014). Dynamic data driven crowd sensing task
assignment. Procedia Computer Science, 29, 1314–1323.
Predic, B., Yan, Z., Eberle, J., Dragan Stojanovic, & Aberer, K. (2013). ExposureSense:
Integrating daily activities with air quality using mobile participatory sensing.
Proceedings of PERCOM workshops (pp. 303–305). .
Quer, G., Masiero, R., Pillonetto, G., et al. (2012). Sensing, compression, and recovery for
WSNs: Sparse signal modeling and monitoring framework. IEEE Transactions on
Wireless Communications, 11(10), 3447–3461.
Rabbi, M., Ali, S., Choudhury, T., & Berke, E. (2011). Passive and in-situ assessment of
mental and physical well-being using mobile sensors. Proceedings of the UbiComp.
369
Computers in Human Behavior 101 (2019) 352–370
Rahim, N., Ahmad, J., Muhammad, K., Sangaiah, A. K., & Baik, S. W. (2018). Privacy-
preserving image retrieval for mobile devices with deep features on the cloud.
Computer Communications. https://doi.org/10.1016/j.comcom.2018.06.001.
Rana, R., Chou, C. T., Bulusu, N., et al. (2015). Ear-phone: A context-aware noise mapping
using smart phones. Pervasive and Mobile Computing, 17, 1–22.
Reddy, S., Estrin, D., et, & Srivastava, M. (2010). Recruitment framework for participa-
tory sensing data collections. International conference on pervasive computing (pp. 138–
155). Springer Berlin Heidelberg.
Reddy, S., Parker, A., Hyman, J., et al. (2007). Image browsing, processing, and clustering
for participatory sensing: Lessons from a DietSense prototype. Proceedings of the 4th
workshop on Embedded networked sensors (pp. 13–17). ACM.
Ruge, L., Altakrouri, B., & Schrader, A. (2013). Soundofthecity-continuous noise mon-
itoring for a healthy city. Pervasive computing and communications workshops
(PERCOM workshops), 2013 IEEE international conference on (pp. 670–675). IEEE.
Sawand, A., Djahel, S., Zhang, Z., et al. (2014). Multidisciplinary approaches to achieving
efficient and trustworthy eHealth monitoring systems. Communications in China
(ICCC), 2014 IEEE/CIC international conference on (pp. 187–192). IEEE.
Schweizer, I., Bärtl, R., Schulz, A., et al. (2011). NoiseMap-real-time participatory noise
maps. Second international workshop on sensing applications on mobile phones.
Seraj, F., Meratnia, N., & Havinga, P. J. M. (2017). An aggregation and visualization
technique for crowd-sourced continuous monitoring of transport infrastructures.
Pervasive computing and communications workshops (PerCom workshops), 2017 IEEE
international conference on (pp. 219–224). IEEE.
Shi, C., Liu, J., Liu, H., et al. (2017). Smart user authentication through actuation of daily
activities leveraging WiFi-enabled IoT. Proceedings of the 18th ACM international
symposium on mobile Ad Hoc networking and computing (pp. 5). ACM.
Shin, M., Cornelius, C., Kapadia, A., et al. (2015). Location privacy for mobile crowd
sensing through population mapping. Sensors, 15(7), 15285–15310.
Sivaraman, V., Carrapetta, J., Hu, K., et al. (2013). HazeWatch: A participatory sensor
system for monitoring air pollution in sydney. Local computer networks workshops
(LCN workshops), 2013 IEEE 38th conference on (pp. 56–64). IEEE.
D. Stojanovic, B. Predic, and N.Stojanovic. Mobile crowd sensing for smart urban mobi-
lity. In: C. Capineri, M. Haklay, H. Huang, V. Antoniou, J. Kettunen, F. Ostermann
and R. Purves, (eds.) European Handbook of crowdsourced geographic information,
pp. 371–382. London: Ubiquity Press.
Sun, J., & Ma, H. (2013). Behavior-based online incentive mechanism for crowd sensing with
budget constraints. arXiv preprint arXiv:1310.5485.
Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few
and how collective wisdom shapes business. Economies, Societies and Nations, 296.
Thiagarajan, A., Biagioni, J., Gerlich, T., et al. (2010). Cooperative transit tracking using
smart-phones. Proceedings of the 8th ACM conference on embedded networked sensor
systems (pp. 85–98). ACM.
To, H., Fan, L., Tran, L., et al. (2016). Real-time task assignment in hyperlocal spatial
crowdsourcing under budget constraints. Pervasive computing and communications
(PerCom), 2016 IEEE international conference on (pp. 1–8). IEEE.
Trossen, D., & Pavel, D. (2007). NORS: An open source platform to facilitate participatory
sensing with mobile phones. Mobile and ubiquitous systems: Networking & services,
2007. MobiQuitous 2007. Fourth annual international conference on (pp. 1–8). IEEE.
Trusted platform module. [Online]. Available: http://www.trustedcomputinggroup.org/
developers/trusted platform module.
Vukovic, M. (2009). Crowdsourcing for enterprises. Services-I, 2009 world conference on
(pp. 686–692). IEEE.
Wakamiya, S., Lee, R., & Sumiya, K. (2012). Crowd-sourced urban life monitoring: Urban
area characterization based crowd behavioral patterns from twitter. In: Proc. Of the
6th IIntl. Conf. On ubiquitous information management and communication.
Wang, Y., Chen, Q., Liu, L., Li, X., Sangaiah, A. K., & Li, K. (2018). Systematic comparison
of power line classification methods from ALS and MLS point cloud data. Remote
Sensing, 10(12), 22.
Wang, D., Kaplan, L., Le, H., & Abdelzaher, T. (2012). On truth discovery in social sen-
sing: A maximum likelihood estimation approach. ACM/IEEE IPSN (pp. 233–244). .
Wang, L., Yang, D., Han, X., et al. (2017). Location privacy-preserving task allocation for
mobile crowdsensing with differential geo-obfuscation. Proceedings of the 26th inter-
national conference on world wide web (pp. 627–636). International World Wide Web
Conferences Steering Committee.
Wang, L., Zhang, D., Pathak, A., et al. (2015). CCS-TA: Quality-guaranteed online task
allocation in compressive crowdsensing. Proceedings of the 2015 ACM international
joint conference on pervasive and ubiquitous computing (pp. 683–694). ACM.
Wan, J., Liu, J., Shao, Z., et al. (2016). Mobile crowd sensing for traffic prediction in
internet of vehicles. Sensors, 16(1), 88.
Wen, H., Hu, Z., Guo, J., et al. (2008). Operational analysis on beijing road network
during the olympic games. Journal of Transportation Systems Engineering and
Information Technology, 8(6), 32–37.
Wirz, M., Strohrmann, C., Patscheider, R., et al. (2011a). Real-time detection and re-
commendation of thermal spots by sensing collective behaviors in paragliding.
Proceedings of 1st international symposium on from digital footprints to social and com-
munity intelligence (pp. 7–12). ACM.
Wirz, M., Strohrmann, C., Patscheider, R., Hilti, F., Gahr, B., Hess, F., et al. (2011b). Real-
time detection and recommendation of thermal spots by sensing collective behaviors
in paragliding. In: Proc. of 1st Intl. symposium on from digital footprints to social and
community intelligence.
Wisniewski, M., Demartini, G., Malatras, A., et al. (2013). Noizcrowd: A crowd-based data
gathering and management system for noise level data. International conference on
mobile web and information systems (pp. 172–186). Berlin, Heidelberg: Springer.
Wu, Y., Wu, Y., Zeng, J., et al. (2017). Pie: A personalized incentive for location-aware
mobile crowd sensing. Computers and communications (ISCC), 2017 IEEE symposium on
(pp. 981–986). IEEE.
D.E. Boubiche et al.
Wu, X., Yang, P., Tang, S., et al. (2015). Privacy preserving RSS map generation for a
crowdsensing network. IEEE Wireless Communications, 22(4), 42–48.
Xiao, M., Wu, J., Huang, L., et al. (2015). Multi-task assignment for crowdsensing in
mobile social networks. Computer communications (INFOCOM), 2015 IEEE conference
on (pp. 2227–2235). IEEE.
Xiong, H., Zhang, D., Chen, G., et al. (2016). iCrowd: Near-optimal task allocation for
piggyback crowdsensing. IEEE Transactions on Mobile Computing, 15(8), 2010–2022.
Xu, J., Fu, J.-R., Yang, D.-J., et al. (2017). Fimi: A constant frugal incentive mechanism for
time window coverage in mobile crowdsensing. Journal of Computer Science and
Technology, 32(5), 919–935.
Xu, L., Hao, X., Lane, N. D., et al. (2015). More with less: Lowering user burden in mobile
crowdsourcing through compressive sensing. Proceedings of the 2015 ACM interna-
tional joint conference on pervasive and ubiquitous computing (pp. 659–670). ACM.
Xu, C., Shen, X., Zhu, L., et al. (2017). A collusion-resistant and privacy-preserving data
aggregation protocol in crowdsensing system. Mobile Information Systems, 2017.
Xu, J., Xiang, J., & Yang, D. (2015). Incentive mechanisms for time window dependent
tasks in mobile crowdsensing. IEEE Transactions on Wireless Communications, 14(11),
6353–6364.
Yadav, K., Chakraborty, D., Soubam, S., et al. (2013). Human sensors: Case-study of open-
ended community sensing in developing regions. Pervasive computing and commu-
nications workshops (PERCOM workshops), 2013 IEEE international conference on (pp.
389–392). IEEE.
Yang, G., He, S., Shi, Z., et al. (2017). Promoting cooperation by the social incentive
mechanism in mobile crowdsensing. IEEE Communications Magazine, 55(3), 86–92.
Yang, D., Xue, G., Fang, X., et al. (2016). Incentive mechanisms for crowdsensing:
Crowdsourcing with smartphones. Biological Cybernetics, 24(3), 1732–1744.
Yang, H. F., Zhang, J., et, & Roe, P. (2011). Using reputation management in participatory
sensing for data classification. Procedia Computer Science, 5, 190–197.
Yan, H., Hua, Q., Zhang, D., et al. (2017). Cloud-assisted mobile crowd sensing for traffic
congestion control. Mobile Networks and Applications1–7.
Ye, M., Yin, P., Lee, W.-C., et al. (2011). Exploiting geographical influence for colla-
borative point-of-interest recommendation. Proceedings of the 34th international ACM
SIGIR conference on research and development in information retrieval. ACM (pp. 325–
334). .
Yu, Z., Feng, Y., Huang, X., et al. (2014). Recommending travel packages based on mobile
crowdsourced data. IEEE Communications Magazine, 52(8), 56–62.
Zappatore, M., Longo, A., Bochicchio, M. A., et al. (2016a). Improving urban noise
monitoring opportunities via mobile crowd-sensing. Smart city 360° (pp. 885–897).
Springer International Publishing.
Zappatore, M., Longo, A., & Bochicchio, M. A. (2016b). Enabling MOOL in acoustics by
mobile crowd-sensing paradigm. Global engineering education conference (EDUCON)
370
Computers in Human Behavior 101 (2019) 352–370
(pp. 733–740). IEEE. IEEE 2016.
Zhang, Z., Choo, K. K. R., Sangaiah, A. K., & Chen, L. (2018). Crowd computing for social
media ecosystems. Applied Soft computinghttps://doi.org/10.1016/j.asoc.2018.02.
045.
Zhang, B., Liu, C. H., Lu, J., et al. (2016a). Privacy-preserving QoI-aware participant
coordination for mobile crowdsourcing. Computer Networks, 101, 29–41.
Zhang, L., Wang, X., Lu, J., et al. (2016b). An efficient privacy preserving data ag-
gregation approach for mobile sensing. Security and Communication Networks, 9(16),
3844–3853.
Zhang, D., Xiong, H., Wang, L., & Chen, G. (2014a). CrowdRecruiter: Selecting partici-
pants for piggyback crowdsensing under probabilistic coverage constraint.
Proceedings of ACM international joint conference on pervasive and ubiquitous computing,
seattle, WA, USA (pp. 703–714). 13–17.
Zhang, D., Xiong, H., Wang, L., et al. (2014b). CrowdRecruiter: Selecting participants for
piggyback crowdsensing under probabilistic coverage constraint. Proceedings of the
2014 ACM international joint conference on pervasive and ubiquitous computing (pp.
703–714). ACM.
Zhang, X., Xue, G., Yu, R., et al. (2017a). Robust incentive tree design for mobile
crowdsensing. Distributed computing systems (ICDCS), 2017 IEEE 37th international
conference on (pp. 458–468). IEEE.
Zhang, X., Xue, G., Yu, R., et al. (2017b). Countermeasures against false-name attacks on
truthful incentive mechanisms for crowdsourcing. IEEE Journal on Selected Areas in
Communications, 35(2), 478–485.
Zhan, Y., Xia, Y., Liu, Y., et al. (2017). Time-sensitive data collection with incentive-aware for
mobile opportunistic crowdsensing. IEEE Transactions on Vehicular Technology.
Zhao, D., Li, X.-Y., et, & Ma, H. (2016). Budget-feasible online incentive mechanisms for
crowdsourcing tasks truthfully. IEEE/ACM Transactions on Networking, 24(2),
647–661.
Zheng, Y., Liu, F., & Hsieh, H.-P. (2013). U-Air: When urban air quality inference meets
big data. Proceedings of the 19th ACM SIGKDD international conference on Knowledge
discovery and data mining (pp. 1436–1444). ACM.
Zheng, Y., Zhang, L., Ma, Z., et al. (2011). Recommending friends and locations based on
individual location history. ACM Transactions on the Web, 5(1), 5.
Zhou, P., Zheng, Y., & Li, M. (2012). How long to wait?: Predicting bus arrival time with
mobile phone based participatory sensing. Proceedings of the 10th international con-
ference on Mobile systems, applications, and services (pp. 379–392). ACM.
Zhu, Y., Li, Z., Zhu, H., et al. (2013). A compressive sensing approach to urban traffic
estimation with probe vehicles. IEEE Transactions on Mobile Computing, 12(11),
2289–2302.
http://opensignal.com/.